Research On Assemblage Detection Technology Of Production Line Nut Based On Machine Vision

The continuous progress of technology and equipment of the hardware performancecontinued to strengthen, the field of industrial production automation degree of continuousimprovement, machine vision inspection technology, as is a new type of technology, beganto applying in many fields, such as industrial, education, medical, remote sensingdetection and intelligent traffic, etc.Nut is a kind of mechanical parts which is widely used in the industrial fields.Especially in some fine production and processing fields, professional precision nut isoften needed to use in production assembly. However, at present, a lot of nut manufacturerstill use the traditional methods of artificial sampling inspection to control the quality oftheir products, which not only heavily depend on human, but also improve the enterprisecost. Therefore, in this paper, the nut detecting system proposed based on machine visioninspection technology has a very important application value to change the currentsituation.Based on the study and reference of the related literature at home and abroad, thispaper applying the Labview image processing toolbox to the circular features of the nut,using the improved Hough transform to extract the boundary and making dimensionalmeasurement to the corresponding geometric features, then epistemic machine accordingto the measurement results to make parts quality decision results. This study on thesimilarity detection has a certain reference value.The main research contents, work, results include the following several aspects:The image pre-processing. The image region of interest extraction to reduceprocessing time and reduce the interference factors. During image pre-processing, wefound that edge was better kept and noise were ideally removed if using median filter,through analyzing and comparing of effectiveness of various slide filters such as averagefilter, median filter and Gaussian filter. As for image segmentation and edge extraction.Threshold Edge extraction based on Sobel algorithm was presented, then small area pixelgroup is eliminated to get accurate edge image.The geometric characteristics of nut detection. Through the analysis of the defective product features and study of the traditional Hough transform and the randomized Houghtransform，the improved Hough transform algorithm of circle detection was presented,which was characterized by quicker speed of iteration，litter stock room and betterprecision of detection. The nut of the edge image using the improved Hough transformalgorithm of circle, compare the result and threshold value to judge nut assembly state...