Research On Vision Detection Technology Testing The RIP Of Conveyor Belt Based On Line Laser Stripe

Belt conveyors are the obligatory equipment for transporting coal and ore etc, thelongitudinal rip of the conveyor accident happens frequently during the working time. Theequipment must be stopped immediately when the longitudinal rip accident occurs;otherwise, the length of the rip will be longer and longer, and the economic expense willbe costly. In order to accomplish longitudinal rip of conveyor belt inspection rapidly andaccurately, this paper study on the questions about the belt tear detection and protection. Inthis paper, a visual detection and protection device has been designed based on the linelaser stripe, to judge whether there was tearing on the belt surface or not by extracting andanglicizing the stripe center. The specific tasks include:A new inspection method for rip of conveyor belt based on line laser stripe isproposed, according to the distortion characteristic of the line laser stripe to detect the belttear. After introducing the line laser, the tear characteristics extraction question wassimplified, the detection efficiency was improved, and the requirements of real-timedetection was meted. The high-contrast features of the line laser stripe, reduced theinterference caused by dust, light and other factors on the image, and the accuracy of thetest result was improved.According to the problem of the accurate segmentation of the line laser stripe, thearea difference between target and background was introduced to the segmentationalgorithm based on statistical criteria, a threshold segmentation algorithm based on thewithin cluster and the area difference between target and background was proposed, thewrong segmentation problem caused by the uneven size of the target (line laser stripe) andbackground of the image was solved. Consider of the intensity distribution of the line laserstripe similar to the Gaussian distribution, a segmentation algorithm combined theiteration method with the searching and approximation method was improved, segmentingthe light image by column, and the segmentation threshold of each column wasindependently of each other, this algorithm can inhibit the interference of the reflectivelight and a variety of noise. The effectiveness of the improved light segmentationalgorithms were verified by experiments. Research on the stripe center extraction algorithm, the initial center of stripe is gotwith the gray barycenter method, the normal direction of the light stripe is calculated onthis basis, and the sub-pixel value on the normal direction is detected by the use ofGaussian curve fitting method. Calculate and analysis the distance of two adjacent pixels,the curvature of the light center, and the migration of the light stripe to fulfill the teardetection. The experimental results verify the effectiveness of each algorithm.Analysis the characteristics of the visual inspection images, use the priori knowledgeduring the segmentation process to improve the efficiency of the segmentation; use morethan one assisted laser stripes to improve the detection accuracy.