Digital image processing-based automatic detection algorithm of cross joint trace and its application in mining roadway excavation practice

This paper proposes a digital image processing-based detection algorithm for cross joint traces of coal roadway heading face. Initially, the acquired images were preprocessed, i.e., adaptive correction was conducted for non-uniform illumination images based on the 2D gamma function. The edge detection algorithm was then applied to extract the edges of the structural plane, followed by the filtration of the non-structural plane noises. Moreover, the Hough transform algorithm was applied to extract the linear edges; finally, the edges were locally connected in accordance with the angle and distance criteria. The experimental results show that this algorithm can be used to reduce the noise caused by non-uniform illumination and avoid the mutual interference of multi-scale edges, so as to effectively extract the traces of the cross joint. Furthermore, Q-system and rock mass rating (RMR), were applied to conduct a quantitative evaluation on the stand-up time of unsupported roof in the four test images. The Q-system quality scores are 26.7, 43.3, 3.1, and 6.7, and the RMR quality scores are 56.84, 58.73, 48.42, and 51.42, respectively. The stand-up time of unsupported roofs with a span of 4.6 m are 30, 36, 7.7 and 14 d, respectively.