The Latitude And Longtitude Detection System Research Base On Machine Vision Technology

With the advancement of technology and social development, Textile market has putforward higher requirements for product quality and equipment performance. The stretchingmachine is an important equipment in the textile machinery, which must be ensured thelatitude location level during cutting the cloth before stretched and stereotyped. Thelevel accuracy of cloth parallels directly determines the quality of the production of cloth.In practice production, the latitude of cloth is often tilted. In the latitude and longitudedetection system of stretching machine, how to quickly and accurately detect thelocation of latitude, and timely correction to make it level, is the key to ensure productquality. Currently, the domestic device mostly uses imported equipment can not meetthe requirements of the textile industry on the micron-level detection accuracy duringto the detection accuracy of latitude location only reaching to millimeter. Therefore,this paper combined with the technology of machine vision and image processing fordetecting the cloth latitude location can reach to micron-level accuracy requirements.The image pre-processing procedure and method firstly are introduced in thispaper. And the three image processing steps of image pre-processing composed ofimage smoothing, image enhancement and edge detection are given. Image processingalgorithms, which are taken for each part, are comparatively studied by numericalsimulation to validate the image processing proper method. Then, feature extraction inlatitude location and longitude detection system is detail analyzed using an improvedangle chain code and cluster-based linear feature detection method. This method firstlydescribed on the edge of the image characteristic angle chain code, extract linearcharacteristic parameters, secondly, image edge may be taken on a straight line,adopting the method of cluster to categorize and judgement,at the same time, derivingthe characteristics of each character space parameters, and positioning and testing theedge of the image. After to analyze the existing priori knowledge, the measured relativeposition of the straight line and the ideal straight line can be seen, and a new set ofimage edge points are gained via the relative location using repositioned edge points.Basing on the total least squares method, a new line can be obtained by the method ofcalculating the char acteristic paramet ers to determine the latitude location parameters. Numerical simulation results show that this approach has a positioningaccuracy, small error and good robustness.According to the performance requirements of the detection system, the mixture oflatitude and longitude based on DSP, FPGA and PC machine hardware detectioncomponent systems are designed in this paper. The main working principle is describedand the part of the hardware circuit modules and software processes are described andanalyzed. Hardware circuit module include: image acquisition module, image processingmodule and PC data transmission module.Finally, experimental results and systematic errors are analyzed, and proposesolutions.