The Development And Research Of Parts Detection System Based On OpenCV And USB Industrial Camera

The rapider development of modern manufacturing industry puts forward higher requirement for the measuring technologies,faster and more accurately.The measurement system based on machine vision has replaced the traditional measurement methods in many places and has been used widely for it has the characteristics like non-contact measurement,high measuring accuracy,fast measurement and so on.The present machine vision part measurement system,taking Card structure as main,lead to some negative results:complicated hardware structure,high cost,long development period,inconvenient maintenance.So the development of the new part detection system based on OpenCV and USB industrial camera was accomplished with Qt development tools and Ubuntu development platform in this paper.The hardwares of this system use the USB industrial camera instead of traditional Cards to simplify the structure of hardware and reduce the costs;And the software developed with Qt and OpenCV realize program modularized and visual programming,which not only reduce the difficulty of soft development and improve procedure operating efficency,but also provide good cross-platform portability which will lay a foundation for developing the embedded machine visual inspection systems.The research work of this paper mainly includes below:1.The shared-streams method proposed in this paper tackles two problems of how the USB industrial camera with non-V4L interface to use the OpenCV resources and how the IplImage structure image to be shown in the Qt control.2.This paper takes advantage of OpenCV calibration method to correct the image and uses the standard parts to calibrate for figuring out the two problems of measuring with variable object distance and calibrating more frequently,.3.Wavelet transform is used to correct the interpolation algorithm image and the handling results are fused with OpenCV image mixing.This method overcomes the shortcomings that the interpolation sub-pixel processing technology has,just like the image is easy to distortion,the image edge is fuzzy or serrated and so on.4.By further researching of each link in the image processing,this paper develops an image processing flow for this system.Combined with OpenCV library function,Qt is used to develop the software of this research.Finally,the measurement precision,stability and reliability of this system are checked through the parts measurement experiment with this software.