Research On Object Detection And Tracking Algorithm And Its Application In Coal Mine

remote video surveillance system. Intelligent visual surveillance technology can minimize manual intervention, reduce the burden on workers, improve the monitoring efficiency. However, affected by complicated mine environmental, noise, dust and uneven illumination, low-resolution factors, target resolution is low. So this paper concentrates to locate, identify and track the objects in complex environment. The main works are as follows:. Object detection system structure was established. Algorithm based on SIFT, using improved K-d tree to search feature points and dimension reduction method to advance system real time ability. Combined RANSAC and L-M nonlinear optimization algorithms to estimate parameters, solved the matching accuracy problems.Proposed a new stitching algorithm to suit for coal mine. Overlapping regions and frames in confusion are checked and ordered by energy distributing of phase correlation impulse function. Multi-band lifting wavelet transform scheme was used to achieve smooth stitching. New algorithm can improve the matching efficiency, and it's suit for coal mine environment.New contour detection algorithm solved the traditional Snake Model has poor anti-noise ability and cannot converge to the concave region in optimizing process. Algorithm improved Snake Model to make it automatically assign the snake points with topology adaptability. Combined GAC , C-V model with level set algorithm for the effective coals segmentation, so it got a better way to solve problems of large coals block the bunker.Based on the Camshift algorithm and fused texture, edge and other features, proposed a new algorithm with adaptive multi-feature fusion and template update. New algorithm established feature template, each feature changed the weight rationally based on different contributions, updated the template adaptively while environment altering.