Nitrogen (N), phosphorus (P) and potassium (K) are not only plant macronutrients but also soil fertilizer indicators, soil N, P and K rapid analysis are of significant importance to agricultural industry. This paper reviewed the recent progress and applications of laser induced breakdown spectroscopy (LIBS) on soil N, P and K detection. To reduce sample diverseness and spectral interferences, samples were mixed, milled, pressed to pellets, and even prepared with a simulated soil matrix of sand. The LIBS instrument was commonly set up with a fundamental frequency laser of 1064 nm and 10-102 mJ, detector usually set to the wavelength of ultraviolet, visible and short-wave near-infrared, collecting spectral at time window of a microsecond scale. Several means were employed to remove air interference and enhance signal quality, including low-energy laser, femtosecond laser, sequenced laser pulse, pre-ablate, optical path optimization, spatial confinement, buffer gases purge and reduced pressure. 746.83 nm was the commonly used N emission line, under the condition of interference avoidance from air N, correlated curves have been used to predict N concentration. Usually used P characteristic line, such as 255 nm, was located at ultraviolet range, methods of linear regression, intensity correction and calibration-free LIBS have been used for soil P analysis. 404.72 and 766.49 nm were the widely used K lines. Beside linear regression, internal standard and calibration-free LIBS, multivariable data mining methods, including partial least squares and support vector machines, were used to analyze soil K. Furthermore, variable selection methods of genetic algorithm, variable importance in projection and the coefficients graph were employed to improve model precision. These researches demonstrates that although challenges remain in terms of matrix effects, spectra and model processing, and instrument development, LIBS is a potential technique for rapid, in suit and multi-elements analysis on soil nutrient.
Potassium detection in the soil is of significant importance for agricultural industry. In this paper, chemometrics methods of artificial neural networks (ANN) and partial least squares (PLS) were comparatively used to detect K in the soil with laser induced breakdown spectroscopy (LIBS). In total, 12 certified reference soils and 17 simulated soil samples with the K concentration of 0.1~3.3% were prepared. LIBS spectra at the wavelength of 723.62~808.24 nm were collected, and then analyzed with ANN and PLS method. The PLS model presented the result of R2val=0.92 and RMSEV=0.26, the ANN model presented the result of R2val=0.82 and RMSEV=0.40. ANN model showed under-fitting and the PLS model performed a better RPD than that of ANN. This demonstrated that the linear PLS model is capable to determinate K concentration in the soil using LIBS.
Objective: Chinese potato staple food strategy clearly pointed out the need to improve potato processing, while the bottleneck of this strategy is technology and equipment of selection of appropriate raw and processed potato. The purpose of this paper is to summarize the advanced raw and processed potato detection methods. Method: According to consult research literatures in the field of image recognition based potato quality detection, including the shape, weight, mechanical damage, germination, greening, black heart, scab potato etc., the development and direction of this field were summarized in this paper. Result: In order to obtain whole potato surface information, the hardware was built by the synchronous of image sensor and conveyor belt to achieve multi-angle images of a single potato. Researches on image recognition of potato shape are popular and mature, including qualitative discrimination on abnormal and sound potato, and even round and oval potato, with the recognition accuracy of more than 83%. Weight is an important indicator for potato grading, and the image classification accuracy presents more than 93%. The image recognition of potato mechanical damage focuses on qualitative identification, with the main affecting factors of damage shape and damage time. The image recognition of potato germination usually uses potato surface image and edge germination point. Both of the qualitative and quantitative detection of green potato have been researched, currently scab and blackheart image recognition need to be operated using the stable detection environment or specific device. The image recognition of processed potato mainly focuses on potato chips, slices and fries, etc. Conclusion: image recognition as a food rapid detection tool have been widely researched on the area of raw and processed potato quality analyses, its technique and equipment have the potential for commercialization in short term, to meet to the strategy demand of development potato as staple food in China.
Wet gluten is a useful quality indicator for wheat, and short wave near infrared spectroscopy (NIRS) is a high performance technique with the advantage of economic rapid and nondestructive test. To study the feasibility of short wave NIRS analyzing wet gluten directly from wheat seed, 54 representative wheat seed samples were collected and scanned by spectrometer. 8 spectral pretreatment method and genetic algorithm (GA) variable selection method were used to optimize analysis. Both quantitative and qualitative model of wet gluten were built by partial least squares regression and discriminate analysis. For quantitative analysis, normalization is the optimized pretreatment method, 17 wet gluten sensitive variables are selected by GA, and GA model performs a better result than that of all variable model, with R2V=0.88, and RMSEV=1.47. For qualitative analysis, automatic weighted least squares baseline is the optimized pretreatment method, all variable models perform better results than those of GA models. The correct classification rates of 3 class of <24%, 24-30%, >30% wet gluten content are 95.45, 84.52, and 90.00%, respectively. The short wave NIRS technique shows potential for both quantitative and qualitative analysis of wet gluten for wheat seed.
Aiming at rapid automatic pest detection based efficient and targeting pesticide application and shooting the trouble of reflectance spectral signal covered and attenuated by the solid plant, the possibility of near infrared spectroscopy (NIRS) detection on cotton bollworm odor is studied. Three cotton bollworm odor samples and 3 blank air gas samples were prepared. Different concentrations of cotton bollworm odor were prepared by mixing the above gas samples, resulting a calibration group of 62 samples and a validation group of 31 samples. Spectral collection system includes light source, optical fiber, sample chamber, spectrometer. Spectra were pretreated by baseline correction, modeled with partial least squares (PLS), and optimized by genetic algorithm (GA) and competitive adaptive reweighted sampling (CARS). Minor counts differences are found among spectra of different cotton bollworm odor concentrations. PLS model of all the variables was built presenting RMSEV of 14 and RV2 of 0.89, its theory basis is insect volatilizes specific odor, including pheromone and allelochemics, which are used for intra-specific and inter-specific communication and could be detected by NIR spectroscopy. 28 sensitive variables are selected by GA, presenting the model performance of RMSEV of 14 and RV2 of 0.90. Comparably, 8 sensitive variables are selected by CARS, presenting the model performance of RMSEV of 13 and RV2 of 0.92. CARS model employs only 1.5% variables presenting smaller error than that of all variable. Odor gas based NIR technique shows the potential for cotton bollworm detection.