Study On Embedded Black-and-white Color-selecting System

The color-selecting system based on machine vision is an intelligent screening device,generally involves optics,image processing,electrical signal control,etc.At present,the image signal processing in the color-selecting system mainly used the computer as control core,thus there exists large volume,low screening efficiency.Today,the rapid development of embedded technology provides new ideas and methods for color-selecting system,and the black-and-white selecting equipment has been widely used for its simple structure and high cost performance.So this paper studies the embedded black-and-white color-selecting system based on machine vision,with the FPGA as the core,which is characterized by good real-time and high reliability,having a broad application prospect in the field of industry and agriculture,etc.Based on the existing black-and-white linear CCD camera as the sensing element,with Xilinx Spartan-3E series XC3S500 E as the processing core,the study completes the design and produce of FPGA main control circuit and the output drive circuit board.The IP core of the FPGA is used to build double-port RAM in its interior to control the reading and storage of the pixel gray value of the material.Since the uneven illumination phenomenon will make the measurement deviated,the paper uses VHDL hardware description language to realize visual data correction,comparative judgment,and delay output drive signal,etc..The output drive circuit controls the open and close of the valves,with the removal of the defective products.With rice as experiment object,and the help of the MATLAB,the research work of various image processing algorithms is carried out and improved,making it suitable for FPGA color-selecting equipment,mainly including: using different filtering algorithm for image denoising,and the processing results of application of edge detection operator are analyzed,and finally improved median filtering algorithm is selected.The image segmentation algorithm is used to get the binary rice image,then the improved morphological processing algorithm is used to remove the small noise and defect points in the image,and smooth the contour of the object.The image connected area marking algorithm is used to identify the seeds in the rice image.At the same time,the paper studies rice appearance detection algorithm,through the contour tracking for rice grain boundary contour,and then calculates the minimum circumscribed rectangle to get the grain length of axis and the size and the shape of obtained information.By setting the appropriate gray threshold value,the paper realizes the test of the chalky rice and mildew.This paper analyzes the characteristics of hardware implementation of different image processing algorithms,and then transfers image processing algorithm to the embedded black-and-white selecting system,using VHDL hardware description language,under the Xilinx's ISE13.4 development tools to write the improved median filtering algorithm,image binarization algorithm,morphology processing algorithm,and the connected component labeling algorithm,and tests the treatment effect and running time of each algorithm,finally realizes rice grain shape parameters and different color detection.