The Research And Development Of Corn Variety Identification System Based On Near Infrared Spectroscopy

Com is one of the main crops in China.As a large country of corn cultivation and consumption,China’s demand for the identification of maize varieties is growing.Using Near Infrared Spectroscopy fast,non-destructive and accurate characteristics to identify corn varieties is an ideal means of detection.In order to realize the corn variety identification system based on near infrared spectroscopy,this paper mainly studied on the corn variety identification method and system realization.Determined the KS method as a sample set partitioning method through analysis and research;To eliminate the noise information in the spectral data,the original spectral data was processed with a smooth+centralized+normalized data preprocessing method;The DPLS+LDA method was used to extract the characteristics of the spectral data information;SVM was the classifier for the model.The correct identification rate of the corn seed single-day data identification model constructed by the above method reached 93.75%.The correct recognition rate of the constructed multi-day data identification model of corn seeds reached 89%,while the PCA+SVM method was only 85.25%and the SIMCA method was 87.5%.The results show that the method proposed in this paper is a method that can better identify corn varieties.The application framework of MFC was the development platfonn of the system.Using OpenCV and TeeChart plug-in to combine database to achieve high speed processing function of the spectral data matrix and the spectral data of the two-dimensional graph display.After testing,the system function was perfect and the identification result was accurate,which can satisfy corn variety identification needs.