Research On Embedded Apple Classification Testing Based On ZYNQ₇010

With the development of computer technology,embedded applications extend to all fields of production and life,and automatic sorting technology is also widely used in the field of agricultural products processing in China.The traditional classification inspection system is cumbersome and inefficient,and the large-scale pipeline inspection system is not suitable for entering the Apple Park for on-site classification inspection.Aiming at the above problems,the thesis designs an embedded apple grading detection system to study the fruit detection grading algorithm and implement it on the FPGA platform.By collecting apple image information,the system selects the YOLOv3-tiny target detection algorithm to isolate the defective apples,and uses the edge detection algorithm to classify the high-quality fruit.Test results show that the system can achieve Apple’s pros and cons detection and size classification,to meet the actual application requirements.The main content of the paper is as follows:1.The thesis uses the Darknet deep learning framework to train the Apple skin defect detection model.The thesis establishes a data set on apple skin defects,and obtains weight values that can identify apples with skin defects through training.2.In order to improve the detection speed of Apple’s skin defects and reduce the memory consumption of embedded devices,design FPGA accelerators for the YOLOv3-tiny network model to achieve the purpose of reducing computing resources and improving computing throughput;3.In order to achieve the size classification of apples,select the edge detection algorithm to extract the edge image of the apple,obtain the maximum diameter of the apple,and then obtain the classification threshold to achieve the classification of apples;4.Design the embedded Apple grading detection system,transplant Linux kernel to the development board ZYNQ₇010,burn the trained YOLOv3-tiny network model weights and edge detection code to the ZYNQ₇010 development board,and perform gcc through the Linux system in the development board Compile to realize the use of FPGA platform for Apple’s skin defect detection and size classification.Based on the ZYNQ₇010 development platform.The paper researched and designed an embedded apple grading inspection system to detect apple skin defects,remove defective apples,and classify the separated and non-defective fruits in size grades.Level classification.