Seed morphological characteristics and weight are important evaluation indicators of seed quality, and they are closely related to seed germination rate and crop yield. In order to detect the quality of melon seeds and improve the efficiency of melon production, a melon seed morphological feature extraction and weight detection system was developed. The light source unit with a ring structure consisted of high power LED lamps, providing stable light for the system; The image acquisition unit collected color images of melon seeds; The weighing unit weighed the weight of melon seeds in real time; The control processing unit processed seed images real-time, triggered the image acquisition unit to collect image and control the collection speed of seeds images, controlled weighing unit to work, and saved seeds image, weight and processing results. The system can extract the number, area, perimeter, long diameter, short diameter and weight of melon seeds. After testing of the system, the detection accuracy of melon seeds number was 100%, detection relative errors of seeds weight, seed perimeter, length and width were less than 5%, and the detection relative error of melon seed area was less than 10%. The results show that the developed melon seed morphological feature extraction and weight detection system can meet the actual needs of melon seed production.
In order to improve the quality of greenhouse vegetable plug seedlings and realize rapid detection of growth information of greenhouse vegetable plug seedlings, a device for detecting the growth information was designed. The detection device contained a weighing unit, an image acquisition unit, a light source unit and a control processing unit, mainly realizing vegetable seedling morphological index of projection area, stem diameter and plant height detection and weight information collection. The light source unit was composed of high-power LEDs. The image acquisition unit was made up of two cameras, the first camera in the vertical downward orientation, was used to obtain projection area parameter of vegetable seedling, the second camera in the horizontal position, was employed to capture of vegetable seedling stem diameter and plant height parameters. The weighing unit adopted a high precision weight sensor to obtain the weight information of the vegetable seedlings. The control processing unit included a single chip microcomputer and a computer. The single chip microcomputer was introduced to control the background board opening and closing, and to control camera work. The computer was mainly used to process images and realize information fusion and the design of humancomputer interaction interface. The software system of the device was developed based on C++ language, including image processing algorithm and control programs. The detection error of the device was less than 5% for morphological indicators and weight information. The results showed that the greenhouse vegetable seedling growth information detection device had high detection accuracy.
Impurity of melon seeds variety will cause reductions of melon production and economic benefits of farmers, this research aimed to adopt spectral technology combined with chemometrics methods to identify melon seeds variety. Melon seeds whose varieties were "Yi Te Bai", "Yi Te Jin", "Jing Mi NO.7", "Jing Mi NO.11" and " Yi Li Sha Bai "were used as research samples. A simple spectral system was developed to collect reflective spectral data of melon seeds, including a light source unit, a spectral data acquisition unit and a data processing unit, the detection wavelength range of this system was 200-1100nm with spectral resolution of 0.14 ~7.7nm. The original reflective spectral data was pre-treated with de-trend (DT), multiple scattering correction (MSC), first derivative (FD), normalization (NOR) and Savitzky-Golay (SG) convolution smoothing methods. Principal Component Analysis (PCA) method was adopted to reduce the dimensions of reflective spectral data and extract principal components. K-nearest neighbour (KNN) and Fisher discriminant analysis (FDA) methods were used to develop discriminant models of melon seeds variety based on PCA. Spectral data pretreatments improved the discriminant effects of KNN and FDA, FDA generated better discriminant results than KNN, both KNN and FDA methods produced discriminant accuracies reaching to 90.0% for validation set. Research results showed that using spectral technology in combination with KNN and FDA modelling methods to identify melon seeds variety was feasible.