A Study On Punched Paper Tape Hole’s Defect Detection Based On Machine Vision Computer

In recent years, with the rapid development of electronics industry, SMD (Surface MountedDevices) components are more widely used in high tech electronic equipment. Chip resistors,capacitors and other passive device are the most popularly used components. Punched paper tape asthe carrier of SMD, the quality of its hole shape directly determines the efficiency of the patchelement packaging effect and the efficiency of surface mount technology. However, theconventional defect detection for punched paper tape is carried out manually. The disadvantage isslow, high error rate and high labor costs. This paper takes the punched paper tape as the researchobject and makes a study about the shape of hole position, defect detection and defect recognition.An algorithm is proposed to detect defect of the punched paper tape by using machine vision. Theexperiments prove that the success rate of the experiment is96%, the defect recognition rate aboutthe variety of hole shape is above90%and the total processing time is40ms. The algorithm candetect various types of punched paper tape, and effectively improve the generating efficiency anddetection accuracy. The main research contents are as follows:(1) According to the characteristics of punched paper tape images,the paper analyzes the imageprocessing technology, puts forward an improved median filtering algorithm, and the algorithm isbetter to save the edge about the hole. According to the processing about tape image filtering, imagesegmentation, edge detection and morphological transform, it obtains the features of real tape, andprovides an effective protection for detection and classification of the holes.(2) The research of punched paper tape about shaped hole positioning and defect detection.According to the characteristics of the hole punch paper tape of the neatly arranges, a large numberof qualified punched paper tape images have been compared and analyzed.And through the shapedhole location and zoning of the shape of the hole tape images,an8-way neighborhood templatematching algorithm has been proposed to detect the defect of the punched paper tape.The algorithmavoids the unnecessary matching operations, saving a lot of time, improve the detection efficiency,and ensures the high accuracy of bidirectional matching detection.(3) Classification research of defects hole about punched paper tapes. Hole-shaped defectclassification has used the BP neural network algorithm. Pass through the use of perforated papertape defect image sample set, and continuous training of the classifier and analysis, the hole-shapedimage recognition neuron weights and the best mix of hidden layer. The result of test shows that, BPneural network algorithm can quickly complete the hole-shaped defect free of high accuracy rate. (4) According to the basic requirements of perforated paper tape pass testing, the overalldesign of the system has been designed, detailing the system’s hardware and software systems. Inthe hardware system, in order to obtain high-quality punch tape image, combined with thehole-shaped defect detection accuracy requirements, the most suitable for the assay of the lightsource, the camera and lens have been selected. In software systems, with a variety software,shaped hole punched paper tape defects can be detected efficiently.