This PDF file contains the front matter associated with SPIE Proceedings Volume 12349 including the Title Page, Copyright information, Table of Contents, and Conference Committee Page.

Real-time and accurate acquisition of precise environmental information for field management is the basis for the realization of precise agricultural operations and intelligent management, and is also an important link in the construction of agricultural informatization. A wireless sensor network-based precise environmental information monitoring system for field management is designed. Collect crop growth environment parameters through accurate sensor nodes distributed in field management, use CC2530 to build ZigBee data transmission and stealing network, and use embedded gateway to complete the information interaction of ZigBee network, WiFi network and GPRS network, and realize remote management of precise environmental information in field management. Monitoring. The test results show that the system has flexible networking mode, accurate data monitoring, wide coverage, and a communication distance of about 350 m, which provides an effective and feasible solution for remote monitoring of field management precision environment.

Better ways of detection of Human Imunodeficiency Viruses (HIV) have been a longstanding goal for scientists. This work proposes an electrochemical biosensor specialized on the detection of reverse transcriptase of HIV which can both ascertain HIV-1 and 2 in one time. The reverse transcriptase has been used in many kinds of detections of HIV, but the biosensor is a new solution. This biosensor is predicted to be ultra-sensitive, relatively low cost, reducing the “window period”, labor saving, and reusable. These predicted data exceeded the most efficient HIV detection method on the market now. By proposing this biosensor, and crafting it in the future, the clinical diagnosis of HIV patients and the exploration into the electrochemical biosensor.

In view of the current situation of environmental monitoring of deep sea aquaculture, the requirements of environmental monitoring of marine aquaculture were analyzed. According to the working requirements and data characteristics of environmental data acquisition sensors, a network model based on wireless sensor network and communication technology was proposed. By constructing a network platform by means of two-level networking, a set of wireless monitoring system for marine aquaculture environment based on ZigBee and GPRS technology has been achieved. The wireless networking was realized by this system through ZigBee dual-direction wireless communication technology. Each water quality monitoring point formed a monitoring network, which converged to a node (gateway) through the ZigBee network. Through GPRS technology and remote communication with the monitoring center server, the real-time monitoring and transmission of marine water quality environmental data, and the processing, storage and management of environmental data were completed in the host computer. As a result, the monitoring and tracking of the deep sea aquaculture environment was achieved. According to the test results, the functional requirements for the collection and management of environmental data were fulfilled by the system.

Stripe rust is a common disease of wheat, and rapid and accurate identification of the damage range of stripe rust at regional scale is of great significance for disease control and prevention. In this study, a new Wheat stripe rust optimized vegetation index (WSROVI) was constructed based on Sentinel-2 remote sensing images and Wheat crown hyperspectral data. By setting a reasonable WSROVI index for threshold segmentation, the incidence area of wheat stripe rust was extracted and the accuracy was verified by ground survey data. The results showed that the user accuracy and mapping accuracy of the spots of stripe rust extracted based on WSROVI were 85.11% and 81.63%, respectively, and the user accuracy and mapping accuracy of the healthy spots were 75.68% and 80.00%, respectively, indicating that the index could be used as a monitoring index of regional stripe rust. This method is simple and efficient, which can provide effective theoretical basis and technical support for regional scale remote sensing monitoring of wheat stripe rust.

In order to solve the problems of feed accumulation, waste of truck space and increasing labor costs caused by single spraying port of traditional silage corn harvester during loading of silage, a safe and effective on-board paving system is designed in this paper; It can realize the average speed and efficient paving of silage corn in the following vehicle harvesting process, improve the vehicle carrying capacity of silage corn feed, reduce the harvesting cost, reduce labor in the silage corn harvesting process, and improve the income of farmers.

Agriculture has been the basic industry of mankind since ancient times. Accurate and robust crop pest identification system is an important step for reliable prediction of agricultural pests in precision agriculture community, which has attracted great attention from many countries. The development of computer vision and deep learning injects new vitality into this field. Many intelligent methods have been introduced into the field of pest recognition. However, long-time model preparation is the bottleneck of these methods, and expensive labels also make this method limited. At the same time, they can not be well compatible and extended when new pests appear. Therefore, we developed a pest identification algorithm based on self supervised learning. In the self supervised learning model(FE-VIT) proposed in this paper, we train a pretraining model through unlabeled data. In the pretext task, the purpose is to make the model match the downstream task(pest recognition). Besides, we simplify the model structure, reduce the time complexity from O (N2) to o (n*k), and make the proposed model Extensible. Our method can achieve about 17% time reduction and 1.1% accuracy improvement. A large number of detailed experiments show the accuracy and reliability of our model. Our code will be released in https://github.com/54zanly/FE-VIT .

Aiming at the verification of the total station commonly used in farmland basic mapping, the traditional method is carried out according to the method of the national verification regulation JJG 703-2003. There are difficulties such as the difficulty of the site selection and construction of the baseline field of the standard, the tedious verification process, the complexity of data post-processing and so on. Based on the analysis of the ranging error of the total station, this paper proposes that the baseline field is transmitted to the high-level total station through the 21 segment combination comparison method, then through the three-stage comparison method, the quantity is transmitted to the verification scheme of the low-level total station. The experimental results show that this method has high verification efficiency, convenient operation, and accurate and reliable verification results.

Time domain reflectometry (TDR) has become the most commonly utilized method for calculating the water content of porous media due to its advantages of rapidity, safety, and non-destructiveness. Nonetheless, the dielectric constant of porous media with the same gravimetric water content varies considerably due to their different porosity, so the TDR method cannot be directly used to measure the gravimetric water content. Although the traditional thermo-gravimetric method has high measurement accuracy, sampling is cumbersome and measurement is time-consuming. Therefore, it is essential to develop an accurate and effective method to calculate the gravimetric water content of porous media. Based on the dielectric measurement results, a third-order polynomial fitting equation describing the relationship between volume water content and apparent dielectric constant of quartz sand is obtained by employing the least square method. Then a TDR method, which can eliminate the influence of gap through secondary compression, is proposed to calculate the gravimetric water content of porous media. The comparative experiment with the thermo-gravimetric method demonstrates that this method has high measurement accuracy, and provides an approach for predicting the gravimetric water content, bulk density, and porosity of porous media, which can be used in field and laboratory applications.

The development of a scalable, low-cost, and versatile sensor platform for the sensitive and rapid detection of heavy metal ions (HMIs) is of great interest to human health and the global ecosystem. Here we report on an all-solution method to fabricate a sensor platform based on nanostructured conductive polyaniline (PANI) in which the electrode materials were functionalized with L-Cysterine (L-Cys). And based on this platform, we realized the simultaneous detection of Pb²⁺ and Cd²⁺ with the detection limit of 11.6 nM and 10.5 nM, respectively. It proves that the proposed sensor has good performance, and broad application prospects in HMIs detection.

The rapid and accurate detection of the quality of wheat kernels, which is the raw material of flour, has been attracting much attention. Some detection methods have been proposed in the last few years. However, these algorithms are incapable of meeting both the requirements of speed and accuracy simultaneously. In order to meet these requirements, we propose a YOLOX_lite model based on the improved YOLOX model. By studying the influence of different color backgrounds on the imaging effect, we design a machine vision system and construct a wheat kernel dataset containing 1746 images of 7844 wheat kernels. The dataset includes moldy damaged wheat kernels, scabbed wheat kernels, sproutdamaged wheat kernels, and undamaged wheat kernels. Using this dataset, we analyze and research the most advanced object detection algorithms. Through optimization operations such as adjusting the CSPDarknet structure and FPN-PAN structure for wheat kernel quality detection. Experimental results show that the proposed method achieves 97.8% accuracy on four kinds of wheat kernels, and the detection speed on the GTX1050 graphics card reaches 49.5FPS, an improvement of 39% to YOLOX_s, which indicates that the YOLOX_lite model meets the requirements of high efficiency, accuracy and reliability applied to the wheat kernel detection system.

Pig behavior can provide information about the barn environment conditions, the adequacy of food and water, health, welfare and productivity. It is of great significance to detect the behavior in time to improve the health level of pigs and economic benefits. This paper reviews the latest technologies for pig behavior detection, and systematically summarizes the characteristics and limitations of some representative pig behavior detection methods in practical applications, including sensor-based pig behavior detection, machine vision-based pig behavior detection, and sound-based pig behavior detection, etc. The next research direction of pig behavior detection is discussed.

Although the concentration of plant hormone abscisic acid (ABA) in crops is very low, it can effectively regulate the growth of crops. Traditional ABA detection methods are limited by expensive instruments and cumbersome sample processing. Therefore, a new method for detecting ABA in ultra-low concentrations is urgently needed. In this paper, a new method for fast and accurate determination of ABA content based on Surface Enhanced Raman Spectroscopy (SERS) was proposed. In this method, a solution of silver-coated gold nanoparticles (Au@Ag) was self-assembled into dense monolayer under the action of capillary gradient force at the air/water interface, which could be transferred to PDMS wafers and filter paper as SERS active substrates. R6G was used as a probe molecule to characterize the Sensitivity and accuracy of PDMS SERS active substrates. When the PDMS active substrate was selected to detect ABA, the detection limit 1×10−11 M was obtained, indicating that PDMS supported core-shell precious metal monolayer substrate can be used as a effective active substrate for detecting ABA. This method is also expected to be applied to the detection of other plant hormones such as corn and soybean.

Alfalfa is known as the "King of Pastures". Only by choosing a suitable planting area can the local climate resources be fully utilized, which is particularly important for large-scale and centralized production. In this research, 443 point records of actual distribution of alfalfa in China and 19 environmental factor variables were selected, based on Maxent model and Arcgis software to simulate the potential distribution of alfalfa nationwide. The results show that the Maxent model has an Area under the Roc Curve(AUC) is 0.821, which is "very accurate". The main factor affecting the growth of alfalfa is temperature. The area suitable for alfalfa growth is 236,000 square kilometers, and the area sub-suitable for alfalfa growth is 584,000 square kilometers. The climatic conditions for growing alfalfa in China have great potential.

By discussing the main factors affecting the agricultural output value of Fujian Province, we can provide countermeasures and suggestions for improving the development level of modern agriculture in Fujian Province. Based on the 2020 statistical yearbook data of Fujian Province, collect and sort out the relevant data of Fujian Province from 2015 to 2020, such as agricultural machinery power, agricultural irrigation area, net amount of agricultural fertilizer application, and crop sowing area, study the impact on the total agricultural output value of Fujian Province, and use the multiple regression model to study and analyze the agricultural output value of Fujian Province, It is found that the main factor of the total agricultural output value in Fujian is the disposable income of the rural population in Fujian Province, and some suggestions are put forward to improve the disposable income of the rural population in Fujian Province.

With the information age coming, the explosion of data is creating opportunities just in the management of the text-documents data in and of itself. Although several information technology is applied in web crawl and information category widely, some vital characteristics of such text classification and natural language processing are not able to obtain directly in specific fields. Therefore, this technical paper propose a multi-head-attention method to classify for Fishery Science and Technology Literature, which aims to help scholars or intelligence agent to obtain the classification results of prepared literature from the database directly. There are mainly three contributions in this paper: (i) the multi-head-attention method and improved training mechanisms is first developed for the text classification research; (ii) a novel compression scheme, referred as PP-MiniLM, based on task-agnostic distillation, pruning and quantization is presented to build the fast, small, accurate classification model; (iii) Furthermore, the experimental results indicate that the prediction accuracy is greatly improved based on the pruning and pooling.

At present, the rapid development of urbanization in China has brought about the relative shrinkage of agriculture. The distribution of cultivated land and population is not balanced, and how to maximize agricultural output value is the main concern. Wind is an important factor in the growth and production of crops. However, in the simulation of wind field analysis, the traditional NetCDF format wind field reanalysis data has too large a time series interval, is relatively single, and cannot be dynamically displayed in a short time series, and the arrow method will cause ambiguity. To solve this problem, based on the 3D virtual earth, this paper proposes a new 2D wind field dynamic visualization method combining the OSGEarth and WRF model, which realizes the short-sequence dynamic visualization simulation of the wind field in the geographical environment and solves the ambiguity problem of the arrow method. It has important guiding significance for mastering the spatial geographic information of agricultural areas and seed dispersal forecasting.

Syzygium album is a plant species with extremely small population and endemic to Yunxiao Country Fujian Province. This supplementary study on biological and ecological characteristics further improved the morphological characteristics of leaves, flowers and fruits and some biological characteristics of Syzygium album, and improved the seed seedling afforestation technology and protective measures. Practice observation records of the three research results show that: The length width radio of blade is about 3:1, the tail tip accounts for 1/5 of the total length of the blade, the dorsal veins of leaves are obviously reticulated at the edge. Autumn shoots germinate every winter solstice, bronze colored annual branches grow eight new leaves, two months from new leaf to mature leaf and leaf color change more than six colors, petiole is bright yellow green. Cymose panicles terminal or axillary, inflorescence axis green tetrarhombic, three to six branches, filament white, flowering period: June-August. Fruiting period: July-November, drupe capsule, the 1000 grain weight of fresh fruit is 119g, one seed, single carpel one chamber, cotyledons two, purple red, mosaic. Seed germination time is not concentrated, seed germinate throughout the year, germinating seeds need to be placed on the plug one by one, replace the nutrient cup according to the seeding height. After careful management for one year, it will grow into strong seedlings with strong roots, after another year of field management only with earth balls can we go up the mountain for afforestation. In summary, Supplementary description of leaves, flowers and fruits of Syzygium album can be used as the basis for interspecific differentiation and identification. Seed seedling raising and afforestation technology can provide practical reference for protecting and expanding the population. Protective measures can provide reference for local protection and ex situ protection.

Rolling bearing is the most central function unit in rotating agricultural mechanical equipment, it is vulnerable to damage in operation because of complex working environment, Therefore bearing fault diagnosis and stateful inspection technology has become research focus of rotary agricultural machines in the present age. entropy theory provides a measure for the complexity and randomness of vibration signals and is widely used in bearing fault diagnosis. This paper summarizes the origin, application and development of entropy theory and the future development direction is prospected.

Diazinon is a highly efficient, highly toxic, low residual organophosphorus pesticide used widely in rice, vegetables, and corn crops. Conventional methods for diazinon detection are limited by expensive instruments and tedious sample pretreatment methods, so a new method for rapid, simple, and reliable trace pesticide residues is needed to ensure the safety of crop products. In this paper, silver nanoparticles were synthesized as active substrates using a reduction technique for SERS signal enhancement. The SERS spectra of diazinon were collected over a wide range of concentrations. The characteristic peaks at 1642 cm^-1 and 1351 cm^-1 were selected for quantitative analysis, and their coefficient of determination (R²) were 0.9929 and 0.9951, respectively. In addition, the molecular structure of diazinon was simulated for the first time, and the vibrational modes corresponding to the characteristic spectra of diazinon were calculated with good agreement using the density flooding theory B3LY P/6-31+G (d, p). These results indicate that the application of SERS to diazinon detection is feasible and has broad application prospects.

Semiconductor UV photodetectors are widely used in civil, military and commercial applications. It has attracted great research interest in recent years. Due to its excellent optical and electrical properties, ZnO has broad application prospects in the field of ultraviolet photodetectors. The optoelectronic properties of single materials are often limited, and composite structures can improve the shortcomings of single materials. In this paper, ZnS materials were composited on ZnO by vacuum vulcanization. The photoelectric properties of PECC UV photodetectors with ZnS/ZnO composite structure as photoanode were also analyzed. The results show that the open-circuit voltage VOC of the UV photodetector with this structure is 0.6 V, the short-circuit current ISC is 5.63 μA, the sensitivity is 401, which is much higher than that of the pure ZnO nanorod array of 119.7, and the responsivity is also increased from 2.81 A/ W increased to 9.38 A/W. The increase in the photoresponse current is attributed to the type II heterostructure formed by the composite, while the slower photoresponse is because the electronic effective mass of the composite structure is larger than that of ZnO with oxygen vacancy defects.

Background. The Shangri-la river valley is a high wine grape growing region for the past decade. Vineyards are found in the Jinsha and Lancang river valleys, with altitudes ranging from 1800 m to 2800 m in this highest wine grape growing region in world. However, the grape berry metabolome of this region are unknown. Methods. Gas chromatography time-of-flight mass spectrometry is a useful tool to identify and qualify metabolites in both the wine grape and wine. Metabolites in the wine grape and wine are responsible for flavour; the flavour differs with different terroir. Secondary metabolism facilitates habitat adaptation, and the wine quality is largely determined by fruit metabolites of wine grapes. Results. The Shangri-la wine-growing region is unique for the grape berry metabolome. Several CS grape berry metabolomics metabolism such as TCA cycle; glycine, serine, and threonine metabolism; glycerolipid metabolism; and amino sugar nucleotide were enhanced in Shandong seashore regions compared with Jinsha and Lancang river valleys. Similarly, lack of carbon fixation, photosynthesis, TCA cycle, and glycine, serine, and threonine metabolism was observed in Shandong seashore as compared with Lancang river valley.The grape metabolomes of the Jinsha river valley were similiar to those of Shandong seashore and differed from those of Lancang river valley due to environment factors such as natural terrior soil types, climate and soil–climate interactions. Discussion. The grape metabolomes of the Jinsha river valley were similar to those of Shandong seashore, whereas they differed from those of the Lancang river valley due to environment factors such as the soil types, climate, and soil–climate interactions of the Shangri-la wine-growing region.The present study provides grape metabolomics data for high-altitude wine-growing regions such as Tibet, Sichuan, and South America.

Melamine is harmful to human health. Excessive use will seriously endanger human health. Therefore, it is necessary to detect the content of melamine in food. In this study, melamine was taken as the research object, and the Raman peaks of melamine were analyzed by density functional theory and compared with the actual samples. After that, melamine was detected by surface-enhanced Raman spectroscopy using gold nanoparticles and silver nanoparticles as substrates respectively. The experimental results showed that the Raman signal of silver nanoparticles as substrates was better than that of gold nanoparticles, therefore, the method has the advantages of simple operation and short process time, and provides a method for detecting melamine.

In the agriculture sector, multispectral camera near-ground observations are currently in demand. Altum multispectral camera is a new type of agricultural multispectral camera, which can be installed on ground observation equipment to obtain near-ground remote sensing images with high resolution and multispectral information. However, the optical system of the Altum camera adopts a multi-lens beam splitting design, and the channel registration problem used near-ground remains to be solved. Therefore, a near-ground image registration method is proposed based on a combination of a channel transformation model and a mapping relationship of the homologous image points. The way first corrects the internal lens parameters using the distortion coefficient of each lens. It then constructs a channel transformation model in the pixel coordinate system based on the transformation relationship between the camera coordinate system and the pixel coordinate system. The following calculates the transformation parameters from the corrected internal parameters and the relative position parameters between the lenses to complete the channel transformation. Finally, the homography model is estimated according to the mapping relationship of the homologous image points to perform inter-channel fine registration. The results show that the error in each band after registration is less than 0.5 pixels when the observation distance is greater than 7 metres. The method can offer a pre-processing service for channel registration for the Altum multispectral camera near-ground applications.

Moths are pests that pose a major threat to food production in China, and the monitoring and prevention of moth infestation is of great significance. To address the problems of a high diversity of moths with minor differences and difficult identification, a semantic segmentation network based on depthwise separable convolution, attention mechanism, pyramid pooling—Depthwise Squeeze-and-Excitation Pyramid network (DSEPNet)—was proposed. The network to extract texture features and wing edge information of moths was enhanced based on the optimization of the model of channel attention mechanism on UNet. The computational speed of the model was increased and the number of parameters of the model was reduced based on the improvement in depthwise separable convolution. A pyramid pooling module was added between the encoder and decoder so that the model could input images of an arbitrary size, while enhancing its ability to learn feature information of different dimensions. DSEPNet was evaluated by ablation and contrast experiments. Compared with UNet, the accuracy, mean intersection over union (mIoU), and F1-Score of DSEPNet were improved by 2.04%, 9.14%, and 4.08%, respectively. Based on the moth dataset, compared with R2AU-Net, the mIoU of DSEPNet was improved by 3.04%. To verify the generalization of the model, comparison experiments were done on the Pascal VOC 2012 dataset. The mIoU of DSEPNet was improved by 0.51% compared with PSPNet and by 0.18% compared with DeepLabv3. Meanwhile, an automatic annotation algorithm for data sets was proposed to solve the time-consuming and laborious process of manual annotation, which can automatically generate semantic segmentation annotation files. DSEPNet can be installed on moth traps to identify moths and monitor the number and species of moths in the area.

Based on High Performance Liquid Chromatography-Diode Array Detection (HPLC-DAD) and chemical pattern recognition, this study constructed 12 batches of polygonatum sibiricum spectral fingerprinting from different areas and sources. Pattern recognition techniques such as similarity evaluation, Principal Component Analysis (PCA) and Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) were used in the research process to screen out the characteristic components as Quality Markers (Q-Markers) of polygonatum sibiricum. A total of 16 common peaks were found in spectral fingerprinting, with similarity evaluation in the range of 0.686 to 0.918. In the PCA analysis, it was found that the three principal components (PC1, PC2 and PC3) characterized the variation information (35.04%, 25.15% and 14.56%) of the original variables, and the cumulative contribution value reached 74.75%. OPLS-DA analysis was used to screen out 6 characteristic components, which are considered potential Q-Markers of polygonatum sibiricum. In this study, the characteristic components were selected as Q-Markers, which provided a reference for the improvement of the quality standard of polygonatum sibiricum.

In recent decade, g-C₃N₄ has been appealing more and more attention because of its ability to directly split water under visible light catalysis. However, it was also greatly limited on account of that it has smaller specific surface area and narrower light absorption range. In this paper, we have developed a specific temperature control program to successfully prepare a kind of reticulated CN, which possessed a fluffy appearance and a porous reticulated microstructure. The BET surface area of the reticulated CN is 33.7 times higher compared with the blocky CN. While under UV-visible light, its hydrogen evolution rate exhibited 67 time higher compared to the blocky CN. Our report will provide bright way to improve photocatalytic performance of g-C₃N₄ for hydrogen decomposed by water.

The main objective of this study was to establish a method for determining the content of orientin, calceolarioside B, kaempferol-3-O-rutinoside in Stauntonia Brachyanthera, and to compare the quality differences of Stauntonia Brachyanthera in different regions. The determination was performed on a Diamonsil®C18 column (4.6 mm × 250 mm, 5 μm) with acetonitrile-0.2% formic acid water as mobile phase, gradient elution, detection wavelength at 340nm, and column at room temperature. Orientin, calceolarioside B, kaempferol-3-O-rutinoside have a good linear relationship with peak area integration values. The regression equation is y = 1700.9x + 2.1473 (R² = 0.9997), y = 1131.5x + 5.8228 (R² = 0.9992), y = 1602x + 2.1948 (R² = 0.9993). The average dosing recovery rates were 98.6% (RSD = 1.76%), 102.6% (RSD = 1.88%), 99.6% (RSD = 1.25%). The High Performance Liquid Chromatography-Diode Array Detection (HPLC-DAD) method is suitable for the rapid determination of three components in the leaves of the Stauntonia Brachyanthera. The three components were different in the leaves of Stauntonia Brachyanthera from different regions, and the comprehensive analysis of Hongjiang has the highest content, which may be related to the picking period, soil quality, climate and other conditions.

The purpose of this study is to study the nondestructive detection technology of pear sugar content based on near-infrared spectroscopy, and to determine the effect of combining different eigenvector extraction methods and model building method on the detection accuracy. First, the near-infrared spectroscopic data and the soluble solid content were collected from 150 pear samples by doing the experiments. The NIR spectral data for pear samples ranged from 833 to 2500nm. Sample sets were divided by using the Kennard-Stone method. Then, the characteristic vectors were extracted by the Principal Component Analysis (PCA) method and the Successive Projections Algorithm (SPA), yielding 14 principal components and 13 characteristic wavelengths, respectively. Finally, ELM prediction models were developed based on the full-spectrum, the PCA-extracted principal components, and the feature variables extracted from the SPA, respectively. The comparison shows the most suitable nondestructive detection model for the pear samples sugar content. The results show that the prediction of the SPA-ELM model is optimal when the sugar content of the pear samples was tested by Nondestructive Testing.

Nowadays, the work of flower classification is relatively less. To solve the practical problems of a large amount of calculation, complex calculation and low classification accuracy in flower classification. This paper adopts the flower image classification method based on transfer learning and visual transformer. Its purpose is to use transfer learning to solve the shortcomings of the transformer model in small sample training, as well as a large amount of calculation of traditional flower image classification. Aiming at the transformer applied by deep learning in image classification, this paper uses two typical models, vision transformer (ViT) and swing transformer (Swin-T) for experimental research. The training accuracy obtained by retraining on small-scale flower image data can reach 96.2% of ViT and 97.6% of Swin-T respectively. The experimental results show that this method has high classification efficiency, good classification effect and strong robustness. Compared with the traditional classification methods, it can greatly improve the accuracy and has a good application prospect in flower image classification.

This study aimed at evaluating the effect of different ultrahigh pressures (UHPs) (non-pressurized, 100, 200, 300, 400, 500 MPa) treatment on nutritional composition and micromorphology of garlic. The results showed that the total sugar content decreased compared with that in non-pressurized, but was not significantly changed among different UHPs treatment (P<0.05); However, the reducing sugar content increased gradually, and reached the highest value at 500 MPa. Both the content of protein and total phenol were significantly higher than that in non-pressure (P⪆0.05); the protein content initially rose and then fell with increasing pressure, and reached the highest at 300 MPa. Change of total phenol content decreased with increasing pressure. The alliin content was all reduced after UHPs treatment. For micromorphology, the starch particles of vascular bundle cells on the surface of garlic were broken and dissolved, and the surface appeared damage indentation after treating with different UHPs compared with the non-pressurized group; but the infrared spectroscopy was shown no difference. The study will provide a reference for UHP to improve the quality of garlic or other foods.

Soil roughness is important for characterizing soil surface, and plays an important role in many land surface models. Previous studies have proposed many methods to describe and model soil roughness. In this study, we use a simple way to measure soil surface roughness and try to find a simpler method to estimate it. We measure soil roughness in a ploughed farmland with an inelastic soft rope as an invention patent says, and then capture the photographs corresponding to the measured field points. Our philosophy is that soil roughness causes the shadows in the photo, so the shadows can be used to invert soil roughness. First, we binarize digital images to calculate the shadow area ratio in each photo. Different shadow area ratios will be obtained with different binarization threshold values, and here we use 6 threshold values to get 6 shadow area ratios. Then we try to use the 6 variables to model soil roughness with random forest method. We divide the dataset to two categories: training data (70%) for modeling and testing data (30%) for validation. We repeat the modeling process for 100 times to get the best model. The model shows a relatively satisfied result with a R2 of 0.81, and a RMSE of 0.05. We conclude that it is a viable way to estimate soil roughness with digital photos. The findings provide a simple and fast means for maximize the potential of digital cameras in estimating soil attributes in the field scales.

The implementation of farmland arrangement project in China has an important role in providing capacity of unified construction and reducing carbon emissions from farmland. The traditional energy saving and carbon reduction assessment and analysis model for farmland arrangement project has mismatched description indexes, and the reconstruction of assessment index system can provide some support for farmland arrangement project, for which the assessment index system is redesigned and an energy saving and carbon reduction assessment and analysis model for farmland arrangement project is established to evaluate the effect of farmland arrangement project. According to the carbon emission in farmland cultivation, the new carbon reduction assessment index system is constructed, the carbon change in the farmland arrangement project is calculated, the weights of each index in the system are determined, the overall assessment process of the model is planned, and the assessment results of the model are finally analyzed.

To mitigate the effects of climate change, we need to reduce greenhouse gases and forests are one component of slowing climate change. I build four models to study forests growth. Model one deals with forest growth. We discussed the prediction model of single tree growth and compare the volume growth models of four kinds of trees find the logistic equation with the highest degree of fitting, give the prediction of average growth of single plant, and obtain the forest growth model through planting density. Model two studies tree carbon sequestration. chose the forest storage conversion factor method to calculate the carbon sequestration of trees. Model three discusses the optimal use of forests. To balance the value of harvested forest products and the value of allowing forests to continue to grow, I propose a relatively good forest management plan. Model four was used to predict carbon sequestration by forests and their products. fit the curve and equation of the residual volume of forest products with time, and find the carbon of a forest and its products within 100 years. The transition point and limit of carbon stock under this model are analyzed.

With the acceleration of economic globalization and the intensification of global climate change, the risks faced by agriculture are also increasing, which has seriously affected the improvement of farmers' living standards and the development of the national economy. In particular, in recent years, the global COVID-19 has had a certain impact on various industries, among which the impact on the agricultural economy presents a short-term diversified trend. How to reduce the impact of agricultural risks on agricultural economy is an urgent problem to be solved in China's agricultural development. Establishing a complete set of perfect agricultural risk management system plays a vital role in reducing agricultural risks, improving economic benefits and the production efficiency. At the same time, it is conducive to improve China's agricultural risk environment, narrowing the gap between urban and rural areas and improving China's agricultural competitiveness. Therefore, this paper discusses the current situation of China's agricultural risks, the causes of agricultural risks and the necessity of establishing an agricultural risk management system. The agricultural information collection system based on RFID technology is proposed, which can effectively improve the ability of farmers to resist risks. Finally, from the formulation and implementation of agricultural protection policies, the improvement of agricultural risk early warning and prevention management system, the establishment of a complete agricultural risk guarantee system, the construction of a futures market in line with China's rural reality improve the dissemination channels of agricultural information and improve the construction of agricultural infrastructure, and put forward suggestions on the construction of China's agricultural risk management system.

After investigation, it was found that there was a temperature difference of 6.09°C before and after the drying area of the small pepper dryer in Guizhou, and the local drying was insufficient. Therefore, this paper starts to optimize and improve the structure of the pepper dryer through simulation analysis, so that the temperature difference between the front and rear of the dryer is within 1.5°C, and the peppers can be dried evenly at the same time. Firstly, this paper studies and analyzes the temperature distribution of the two vertical baffles, and the scheme still has a temperature difference of 3.54°C. Secondly, arc and horizontal slab hybrid optimization scheme is analyzed, and the maximum temperature difference is 1.44°C, so the temperature distribution is very uniform and the temperature difference is small, which can fully meet the actual drying requirements.

Rural environmental protection and governance is an important link for the sustainable development of township economy, the sustainable utilization of resources and the realization of short-term and long-term interests. Many villages in Western China show the characteristics of many points and a wide range. The objects of rural environmental protection and treatment are mainly domestic waste and sewage. In order to save resources and reasonably set the site selection of environmental protection facilities, it is necessary to obtain a large number of rural topographic maps in advance. However, there are many villages in Northwest China, and traditional means can not obtain drawings and data quickly and accurately. UAV aerial survey has high speed, high precision and high operation efficiency, but the cost of industrial UAV is high. Thanks to the development of sensor technology, the non measurement camera pixels carried by consumer UAV basically meet the industry requirements. With the help of phantom 4 Pro UAV platform, this paper has completed the rural topographic map mapping of a certain area in the northwest, which has passed the verification, its accuracy fully meets the design requirements of facility location, and provides technical support for scattered regional aerial survey.

The growing area of kiwifruit in China has resulted in an annual increase in labor costs during the harvest period. In this paper, we present a dual-arm picking robot, which is specially designed for picking kiwifruit in trellis orchards to meet the needs of labor shortages. The performance improvement is reflected in the use of deep learning to identify on-site kiwifruit and formulate picking strategies to achieve the purpose of efficient picking. The harvesting robot consists of two robotic arms and a mobile platform. The end of the robotic arm is equipped with an end effector to ensure the safe picking of kiwifruit. The vision system consists of a deep learning system and an RGBD camera, and uses yoloV4 model to identify kiwifruit. The performance of the harvesting robot is evaluated by the test, and the results show that the harvesting success rate of the robot is 90%, and the target recognition rate is 95%.

Agricultural big data refers to an expert system of soil data mass storage platform based on Hadoop, which is formed after combining the characteristics of agricultural rationality and seasonality. Using big data technology software, aiming at the problems of massive growth and low sharing of current agricultural soil chemical characteristic data, and combined with Hadoop framework. In order to realize the platform, HBase + HDFS is used to build a cloud storage platform; Aiming at the problem of server load balancing, partition strategy and dynamic load balancing strategy are proposed; Parallel clustering is introduced to realize the classification and mining of soil chemical characteristic data; Finally, the storage platform is verified. The results show that the platform shows excellent performance in performance and clustering result scheme, which has great reference value for the application of big data analysis technology in agricultural production. In building a new countryside, the application prospect of big data technology is broad.

Orchard mechanization is the way to improve orchard productivity. At this stage, the mechanization level of our domestic orchard demonstration base is constantly improving. This article summarized the development process of mechanization of demonstration orchards at home and abroad. The author referred to the orchard demonstration bases at home and abroad, finally taking Guangdong province as an example and summed up four important suggestions for the mechanization development of demonstration orchard bases.

Seedling emergence rate is one of the important factors affecting crops yield. In this paper, the estimation method of peanut seedling emergence rate based on UAV visible light image was proposed to improve the timeliness and accuracy of peanut seedling emergence rate information acquisition. Firstly, obtained high resolution remote sensing images of June 25, 2021 and July 3, 2021 by self-built low altitude unmanned aerial vehicle platform at summer sowing peanut seedling stage, and the number of peanut plants was identified and counted by combining HSV and Otsu adaptive threshold method. Then, combined with visual interpretation data, linear regression was used to establish a linear relationship model between visual interpretation and computer automatic extraction of peanut plant number. Finally, the information of peanut emergence rate was extracted based on this model to evaluate the sowing effect and quality. The results showed that the correlation coefficient R2 between computer automatic identification of plant number and artificial visual interpretation of plant number was more than 0.85 in two periods. The overall accuracy of peanut plant number extraction was higher than 95 %, and the average relative errors of seedling rate estimation in two stages were 6.0 % and 0.4 %, respectively. The monitoring accuracy of UAV in the middle and late stages of seedling was better than that in the early stage of seedling, which was suitable for peanut seedling rate monitoring. The emergence rate of narrow high ridge sowing was better than wide high ridge sowing. This paper quantitatively analyzed the sowing effect and quality of wide high ridge planting and narrow high ridge planting, and provided technical support for fine management of peanut industrial park.

The automatic navigation control system of the intelligent electric tractor is mainly studied in this research. Based on the analysis of the overall scheme of the control system of the electric tractor, a path tracking algorithm is proposed for the automatic navigation control system, and the working principle and control process of the algorithm are analyzed in detail. Finally, the effectiveness and stability of the automatic navigation control system are verified by a tractor automatic navigation experiment in a field.

Nowadays, China's agricultural development has entered a new stage of transformation from traditional agriculture to highstandard farmland construction. The development of intelligent, high-standard and water-saving agriculture has become the development direction of modern agriculture. However, although my country’s agricultural sector has achieved a certain degree of mechanization, there are still pain points that are not water-saving, unintelligent, and non-interactive. It is urgent to strengthen the promotion and application of today’s advanced technologies in the agricultural field to achieve refined agricultural production management. To achieve the purpose of rational use of agricultural resources, reducing production costs, improving the ecological environment, and improving the quality and quality of agricultural products. Taking wheat planting as an example, this article focuses on the construction of digital twin farms. First, establish a health monitoring infnsor networks to collect farmland information and facilities operation. The rapid identification of the water; combined with meteorological and soil moisture data, the use of deep learning algorithms to build an information processing system based on multi-source data fusion, build a wheat digital twin information model, and carry out irrigation risk prediction; finally apply irrigation strategies on the wheat entity, The information model of wheat and the physical model are interacted, and the purpose of intelligent water-saving irrigation can be realized through multi-mode irrigation and fertilization control.

It is difficult to distinguish different maize varieties accurately by naked eyes or simple instruments because of their similar appearance. However near-infrared spectroscopy (NIRS) provides a quick, accurate and non-destructive way to analyze of maize varieties. Near-infrared spectral images of four types of maize seeds are collected in the wavelength range of 1100nm-1900nm. The near-infrared spectra are preprocessed by standard normalization transform and wavelength preselection. Eighty near-infrared spectra are obtained for each type of maize seeds, 60 of which are randomly selected as the training set and the remaining 20 spectra as the testing set. Then, support vector machine (SVM), SVM-Genetic algorithm (SVM-GA), backpropagation (BP) neural network, and principal component analysis-K nearest neighbor algorithm (PCA-KNN) are established for the variety identification. The average accurate identification rates are 93.75%, 97.50%, 92.50% and 87.50% respectively. The SVM-GA model shows better performance for near-infrared spectra of maize seeds, and the classification accuracy increases nearly by 4%. This work provides an effective method for rapid automatic classification of maize varieties.

Intelligentization of farming is a research hotspot. In this paper, we present several models that use collected data to predict the temperature and growth status of individual pigs. As far as we know, previous research in this field has focused on data acquisition, and relatively little research has been done on data analysis. Our work takes a pig farm as an example and uses statistical learning and artificial intelligence to make better use of this data.

As a local variety in Dazhou, crispy red plum has obvious regional characteristics, and its ripe fruit is still quite hard. Crispy red plums are respiratory variable fruits. In a short period of time after harvest, the respiratory intensity rise sharply, and the quality of fruits declines, resulting in difficulties in storage. Shelf period, sales period and processing period cannot be guaranteed. Difficult storage and processing have become the core factors hindering the development of the crisp red plum industry. This paper discusses the effects of different temperatures on post-harvesting crisp red plums under intelligent air conditioning storage. By measuring the changes in the antioxidant enzyme activity of fruits during the storage of crisp red plums, the mechanism of different temperature treatments adjusting the physiological activity of crisp red plums is revealed. The results show that low-temperature storage can delay the arrival of respiratory peaks and reduce respiratory peaks during plum fruit storage. At the same time, it can inhibit the increase of relative conductivity and the generation of MDA during the storage of crisp red plum fruit. In addition, low-temperature treatment can also control the reduction of SOD and POD activity during the storage of plum fruit, ensure the normal operation of the reactive oxygen system of fruit removal, and reduce the damage caused by oxidation stress to plum fruit.

The rapid development of "Internet +" agriculture has spawned a new form of modern agricultural development— agricultural virtual agglomeration. Empirical methods such as DEA-HP and OLS regression were used to explore the degree and influencing factors of modern agricultural virtual agglomeration. Based on the conclusion, some countermeasures and suggestions are put forward, such as perfecting the management system of agricultural virtual agglomeration, improving agricultural information infrastructure and technical level, and cultivating technical talents.

As one of the key technologies for pig behavior recognition, multi-target tracking is of great significance in the field of automated monitoring of large-scale pig farms. Aiming at adhesion and occlusion problems between pigs, this paper proposes an improved IOU-Tracker pig tracking algorithm suitable for real breeding scenarios, and combined with the YOLOv5s network to achieve real-time pig tracking. In the tracking algorithm, the strategy of preferential matching between the Narrow Rectangle of IOU and the Priority Matching by IOU Value to solve the adhesion problem of pigs; it integrates an algorithm of Location Prediction in several frames to predict the reproduced position of the lost target within a certain number of frames, effectively reducing the change of the target ID due to factors such as occlusion and detection frame jitter. Construct pig tracking datasets under different perspectives and scenarios, a total of 7 pig videos, and use the MOT benchmark as the multi-target tracking evaluation standard, also verify on the public dataset of MOT16 and the self-built pig data set. Experiments show that on the public data set MOT16, the algorithm in this paper has reached 48.57% and 88.67% on the MOTA and MOTP indicators respectively; the IDSW times are only 297, which is 59.3% and 49.3% lower than the IOU-Tracker and DeepSort algorithms respectively; On the self-built pig data set, the MOTA index of the algorithm in this paper reached 97.60%, which is 0.37% and 9.39% higher than the IOU-Tracker and DeepSort algorithms respectively; MOTP reaches 98.38%, which is 0.1% and 2.2% higher than IOU-Tracker and DeepSort algorithms respectively; IDSW times are 40 times and 34 times lower than IOU-Tracker and DeepSort respectively; And the FPS reaches 33, which can meet the real-time tracking requirements. Therefore, the algorithm in this paper has a good effect on pig tracking in pig farms, and provides a technical basis for the automatic monitoring of large-scale farms.

In recent years, drones have been widely used in agriculture due to their simple operation, low cost, and good maneuverability, which is used to monitor crop yield and climatic conditions. However, due to the limited capacity of the lithium battery it carries, the airborne time and operating area of drones are limited, so a fast and convenient charging method has become a key issue on the development of drones. The current charging methods are divided into plug-in charging and wireless power transmission (WPT), the traditional plug-in charging is gradually replaced by WPT due to the easy wear and tear of lines. While WPT has great application value for the power supply of drones due to its advantages of high transmission efficiency and safety. In view of this, this paper proposes a WPT system based on transmitter composite topology switching for agricultural drones applications. Firstly, the constant current (CC) characteristics and zero phase angle (ZPA) operation of SP/S topology are discussed through theoretical analysis. Secondly, the CC and constant voltage (CV) output characteristics of the WPT system based on the composite topology switching on the transmitter side are analyzed, and the parameter design method is proposed. Finally, the circuit simulation is carried out by the software MATLAB-Simulink to verify the rationality of the proposed system.

Leaf chlorophyll content is one of the key indices for plant growth state and nutrient management. This study was designed to develop a new prediction model for tobacco leaf chlorophyll content using Unmanned aerial vehicles (UAV)-based Multispectral Camera. An experimental field with four levels of N application rate (67.5, 142.5, 217.5, 292.5 kg N/ha) was conducted. Five bands (450, 560, 650, 730, 840nm) spectral information and destructive measurements of leaf area, leaf dry weight, leaf chlorophyll content were determined from representative upper, middle and lower leaves at leaf maturity stage. 58 vegetation indexes were calculated and regressed against the measured leaf chlorophyll content using stepwise regression analysis (SR), partial least squares regression (PLSR), random forest (RF), and Artificial neural network (ANN). The chlorophyll content per unit area ranged from 0.73 to 5.15 mg/cm², and the chlorophyll content per unit weight ranged from 0.14 to 0.91 mg/g. The relationship between chlorophyll content and N rate was directly proportionally. The leaf reflectance at different N levels was basically the same, leaf showed lower reflectance at 450 nm, 560 nm, 650 nm and high reflectance at 730 nm and 840 nm. Signal band reflectance has a lower correlation (|r|<0.75) with the chlorophyll content. The models’ R2 for predicting chlorophyll content per unit area ranged from 0.60 to 0.88, which has more accurate than the model predicting chlorophyll content per unit weight. The ANN model exhibited better performance for predicting chlorophyll content, with R²=0.875 and 0.743. These findings have important implications for improving tobacco growth-related traits in precision agriculture.

Focusing on the development trend of maize breeding technology, this paper uses the methods of bibliometric analysis, co-word network analysis, visual analysis and content analysis to retrieve the papers and patents in CNKI literature database and incopat patent database, uses Excel, network analysis integration software, visual analysis system to from the aspects of article load, high-frequency authors, research hotspots, core journals, patent layout, Carry out in-depth mining and visual analysis of research situation in this field, and prospects for the future research. The results show that the number of literatures on maize breeding technology has increased year by year since 1957, and researchers' attention to this field has gradually increased. Under the strategy of fighting a turnaround in the seed industry, it will continue to receive attention. Relevant research is mainly focused on crop disciplines, and the fund support obtained is relatively diversified, basic and applied basic research institutes account for the largest proportion, and more than 50% of the literature is mainly published in professional journals, Patent applications are mainly applied by scientific research institutions and enterprises, supplemented by colleges and universities. Research hotspots mainly focus on maize genetic breeding, biological breeding, waxy maize, sweet maize, breeding technology, disease resistance breeding. In the future, it will be a research hotspot in the aspects of ‘machine harvest + transgenic’, industrialization of biological breeding and digitization of maize production.

Based on the high-throughput of a company in Luoyang Chinese aluminum casting and rolling imported belt production line, and the forming property of AA5052 alloy on the production line were analyzed by mechanical tensile test and microscopic analysis methods, comparative study and blooming process and hot rolling; at different temperatures of AA5052 cold rolled plate was annealed. To explore the process of AA5052 alloy under the law, and the development of the annealing process for high volume continuous casting and rolling production line of products; also from the micro perspective on the AA5052-H32 sheet forming properties especially the bending of instability phenomenon of the causes, the rolling and annealing process of different forming and the performance of the organization is optimized, finally formed a set of suitable for high-throughput production of AA5052-H32 alloy under continuous casting and rolling and annealing system.

Agricultural pest target detection is one of the important tools for agricultural pest control. Aiming at the problem that the traditional pest detection model is affected by factors such as the large variety of pests and the large variation of instance scale, resulting in low detection accuracy, a detection network KPL-YOLOv5 that strengthens key point information is proposed. Redesign the core module CA-Block in the backbone network so that the network adaptively extracts multi-scale instance features and selectively filters redundant gradient information. The key point localization fusion (KLF) module is proposed in the regression prediction network to enhance the mining and fusion of coordinate information of multi-scale targets and improve the localization bias problem in the regression prediction stage. Experiments show that the proposed method improves mAP by 4.08% compared to the YOLOv5 algorithm on the study dataset, and improves compared to a variety of mainstream algorithms, enabling high-precision localization and detection tasks of agricultural pests in complex backgrounds.

In order to improve the utilization rate of water resources and realize the intelligent irrigation of water and fertilizer integration in greenhouses, a multi-sensor dynamic monitoring network was built around the crop growth environment, and an intelligent irrigation system based on STM32 microprocessor was designed. On the basis of fully considering the key factors of the crop growth environment, according to the real-time data, the type and proportion of fertilizer required by crops are analyzed and calculated. The model can intelligently control the amount of water and liquid fertilizer based on the different growth periods of crops to achieve precise water and fertilizer control. By verifying the hardware and algorithm, the results show that the method proposed in this paper can not only save water, but also improve the quality of crop growth.