The Detection And 3-D Model Construction Of Irregular Void Profile

As the rapid development of computer technology, artificial intelligence and thinking scientific research since the 1970s, digital image processing technology has grown up into a higher and deeper level, and has been widely used in fields of image communication, office automation system, geographic information system, medical equipment, and others. It could help improving the effectiveness and accuracy of the operation by researching the method of detecting inner structure of irregular void.This paper studies a detecting method of irregular void profile based on machine vision, and focuses on detecting the shape and size of the cross-section profile of the void. The research of the software system is the key point, including image acquisition and display, image pre-processing, object profile extraction, object profile fitting, as well as the three-dimensional reconstruction. Object profile extraction is the most critical part of the system. Firstly, the methods of image filtering, image binary, profile thinning and other pre-processing methods are adopted in consideration of the characteristics of the experimental images, and results show that an anticipant effect is achieved. Secondly, as the influence of the light or other factors, the contour line might be incomplete, so a profile tracking algorithm based on curvature vector was proposed. Every segment is accurately extracted in clockwise order, and then several contour segments are arranged in the correct sequence. A curve fitting method is used in break points. Finally, the three-dimensional reconstruction model is obtained on the basis of processing a series of images.In this paper, a software system is designed according to the image processing of irregular voids profile. Then the software system is tested, and results show that the accurate profile model of void is obtained. This paper studies the key technologies of detecting the size and shape of inner cross-section profile of irregular void in unstructured geographical environments. It could supply users with necessary geographical information, especially the information of underground irregular voids or the unknown voids, such as the ruins caused by mine collapses, earthquakes or other disasters, and improve the efficiency, security and accuracy of operations.