Machine Vision Technology And Application Of Line Detection

Digital image processing technology has developed rapidly, based on machine vision applications in the industrial field and gradually toward the practical stage. Using the theories and algorithms of image process to resolve the problems in industry and real life is a significative work. That is also the destination of image process. But nowadays, machine vision systems are rarely applied in factories. One of the reasons is the restriction of the algorithms or theories of image process.Detecting the assigned images is one of the most frequently tasks in factories. And detecting the assigned the straight lines is one of suck kind of problem. How quickly and effectively detect the number of range, fixed line is discussed in this paper. This paper detailed analysis the existing algorithms, and presents the weighted least squares fitting method based on a linear gray scale information. The tick as experimental subjects, the results show that the algorithm is high accuracy compared to traditional methods. Finally, the algorithm is applied to pointer meter reading system. Major research achievements and Innovations are as follows:1. This paper describes some common line detection methods and improved straight line detection methods, review the principle ideas of each method. Finally, four methods will be used for the detection of line dial scale, comprehensive analysis the effect of each line detection methods and advantages and disadvantages.2. In order to resolve the problem of low accuracy in processing binary images, and comprehensive analysis the principles and disadvantages of the existing algorithms, this paper presents the weighted least squares fitting method based on a linear gray scale information. This method can improve the detection accuracy of the line.3. Based on the performance of this algorithm is the main factor——the selection of weights, the paper analyzes a method of selecting weights, so as to enhance the performance of the algorithm in detecting lines in complex images with high efficiency.4. In order to improve the accuracy of line detection in the grayscale, this paper presents linear feature detection method base on sub-pixel. This method will improve the accuracy of line detection to the sub-pixel level.The system simulation platform programming using VC6.0. Extraction of tick this step uses the above methods, we can get the exact location of the scale line. In this way, It can lay a good foundation of reading in the system, and it has some practical value.