Face Recognition In Personnel Attendance System Of Coal Mine

The accuracy of coal mine attendance system can not only meet the production and management in the field of mine safety, but also provide a timely basis for making mine accident rescue decision and avoid huge losses of state property and lives of the miners. Therefore, an efficient attendance system plays an important role for coal mine safety Taking the presence of coal dust, dust, water contamination into account, underground staff palms and fingers are susceptible to contamination, so usually hand patterns, palm prints, voice prints and other biometric techniques can not be applied to mine attendance system. The mine workers face, although contaminated, the outline of the face and features change little relatively, so the intuitively non-contact face recognition technology of underground staff attendance system has become the topic research of coal attendance technology. This paper studied on face recognition problems in the coal mine face attendance technology, and the following works are included:We studied ASM algorithm research, which is based on active shape model, for face detection and face tracking in coal mine complex environment. The algorithm combined with human facial geometry, by the method of translation, stretching, rotating, adjusting the position of face in the overall image, after learning local deformation in the training set, established a linear shape model and the patch model of the face. The algorithm solved face detection and face tracking problem of underground staff in the coal mine attendance system.Aiming at the difficulty in identifying face of coal mine workers accurately, the paper based on bi-dimensional principal component analysis 2D2 PCA algorithm, introduced block processing idea to reduce the dimension of feature vectors, and gave a face recognition algorithm which combined the two-dimensional discrete cosine with block 2D2 PCA, which could improve the recognition results and recognition speed of PCA algorithm effectively. To further address the recognition accuracy and speed of the coal mine personnel attendance system, the article gave a face recognition algorithm based on SURF and bi FLANN match method. Firstly, the rapid Hessian matrix in SURF algorithm is used to detect facial feature points, generating SURF feature descriptor; then we combined positive and negative of matrix trace of Hessian with two-way FLANN matching search algorithm, excluding the mismatching points in the matching recognition, by matching the image SURF descriptors accurate face recognition. By comparison with the experimental results of LBP, SIFT, previous SURF algorithms, it showed that the revised algorithm can improve the SURF algorithm recognition rate and accuracy effectively, and lay the foundation for face recognition in coal mine attendance system.The paper studied the corresponding algorithms on mine workers face detection, tracking and identification three stages respectively, initially solved algorithm issues of coal mine attendance system based on face recognition in the complex environment and laid a technology foundation for the further development of face attendance system.