Research On Surface On-line Detection System Of Plastic Connectors Based On DSP

Plastic connectors are widely used in construction,machinery,chemical industry.Some appearance defects can affect the function of the product,so the detection for this product is an indispensable link in production.Compared with artificial detection,machine vision has higher speed andprecision.Therefore,machine vision is used widely in industion production filed.The thesis combines a cross subject, developing a on-line detection system of the plastic connectors in order to sovel the bottleneck of the artificial inspection.This thesis firstly discussed the background and significance of on-line detection.Then,accessing to domestic and internation research literature and to make the overall design scheme and technical route.Next,in order to provides the theory support of improvement of detection algorithm, this thesis introduces the basisi of digital image processing and some common visual detection algorithm. Then according to the system’s design requirements, the thesis make a design task and complete the hardware system design and selection based on DM6437visual detection platform. Later, according to the process of the industrial field detection,the thesis make a software design,which consist of image acquisition, image detection and product separation three modules. Among them, the image acquisition module includes the configuration of video decoding chip TVP5150and video processing front and back end.Image detection module is consist of image filtering,template matching and image segmentation.Based on reading and testing,the thesis improve a median filtering algorithm.As the speed of traditional template matching is slow,the thesis proposes a fast matching algorithm based on integral image and pyramid search strategy method.On the basis of analysis,the thesis puts forward a kind of high efficient and practical edge measuring method, which used to detect the integrity of the edge of the products. And product separation module is completed through the communication between DM6437and omron PLC,which used is Hostlink communication protocol.Finally,by using DSP optimization technology to optimize the software code and designed the experiments to analyze the experiment results and the factor of error.This inspection system can satisfy the speed and precision of detection,also reached the index requirement,and stability practicability is good.