Research on product design and application of line array intelligent camera based on FPGA and SSA-SVM algorithm
To improve the detection accuracy of intelligent cameras for grain quality detection,this study proposes a linear array intelligent camera that collects grain information through TCD2564 sensors and builds intelligent algorithms through FPGA.In the im-age preprocessing section of the camera,the discrete Gaussian filtering algorithm after template adjustment is used for image denois-ing,and the two point correction method and time delay method are used for grayscale correction and color misalignment correction of the image;In the analysis of grain quality,the improved maximum inter class variance method is used for target extraction,and SSA-SVM is used to classify the quality of grains based on their color,texture,contour,and other features.Through experiments,it has been proven that the template adjusted discrete Gaussian filtering algorithm has better image denoising performance,and the two point correction method and time delay method can effectively correct image grayscale distortion and color misalignment;The improved max-imum inter class variance method has better image extraction performance.SSA-SVM achieves an accuracy of 98.34％for the quality classification of peanuts,with a single division time of 2.747 seconds,indicating good grain quality classification performance;This linear array intelligent camera has high quality detection accuracy for grains of different volumes,and even reaches 98.55％for medi-um sized grains such as soybeans.It has certain application value in automatic detection of grain quality.
linear array sensorFPGAimage preprocessingmaximum inter class variance methodSSA-SVM
万方数据
维普
