Dynamic Potato External Quality Based On Machine Vision Non-destructive Testing

In recent years,machine vision technology has developed rapidly,and it also plays an irreplaceable role in defect detection.In order to increase the added value of potatoes,increase the competitiveness of the potato market,and promote the development of the potato industry,machine vision-based detection technology is used to classify potatoes.In the process show their talents.This article mainly from the following five aspects of the study:(1)In order to extract the characteristics of potato accurately and effectively,image smoothing and image preprocessing techniques such as image segmentation techniques,image erosion and expansion,opening and closing operations,image background operations,and minimum external moments were studied.(2)The defect detection module compares the advantages and disadvantages of the SUSAN operator and the gray level threshold method,uses the gray level threshold value of 0.3 to extract the defective part,only processes the defective part image,and reduces the number of spots by passing the area threshold value of 200.The misjudgment of potato increases the accuracy of the recognition of wormholes and mechanical damage.(3)The weight detection module,through the extraction of the geometric parameters of the potato,compared the three detection methods,innovatively proposed a three-factor area parameter method for fitting,the correlation coefficient was 0.9824,has a high correlation,and in the actual detection During the process,the problem of individual potato errors was improved and the overall detection accuracy was also improved.(4)The potato shape detection module obtains seven moment invariant parameters of potato as input parameters of BP neural network,and establishes a 7×5×1 neural network model suitable for potato shape classification.(5)Use matlab parfor instead of for loop to perform synchronous operations.Through testing,parallel processing improves the speed of the whole classification to a certain extent,and the more potatoes are detected,the more obvious the effect is.The system uses the matlab GUI graphical interface as a development platform to design a hierarchical system software window to achieve basic operating functions.The potato non-destructive testing device described in this paper is of great significance for the innovation of the potato classifier,and has certain reference value for the application of machine vision technology in the actual production and processing.