Coal Mine Personnel Attendance System In Face Detection

Accurate underground personnel attendance information not only can satisfy the needs of production and management in the field of coal mine safety, but also provide decision-making basis for the coal mine accident rescue in time. With further increasing supervision efforts for the safety of underground personnel in our country, the coal mine’s contamination has used a variety of biometric identification technologies for underground personnel attendance. Taking into account the underground personnel in the attendance, for the reason of the dimly light in the underground tunnel, or parts of body attached by the black dust, or some people congenital absence or pollution of fingerprints, or strict poses demanding by iris attendance and other factors, intuitive and non-contact face recognition technology is more suitable for the needs of underground personnel attendance. In this paper, when the underground personnel are out of the coal well, the face will be partially obscured. Our work in face recognition technology has been focused on:Face contour extraction and dynamic detection.Due to underground personnel’s contamination by pulverized coal, the facial contours are often affected by the facial skin color surrounding it, which leads to facial contours’ unclear. When the facial contour is extracted by the traditional Snake model, the skin surrounding face has an obvious reaction to face contour extraction, while the traditional ASM model is also more sensitive to facial contour’s light and noise. This article will organically combine Snake model and the ASM model, and propose a face contour extraction method to contaminate face. Firstly, the ASM model will be used to locate the broad facial contours of underground personnel, and then the Snake model will be applied to evaluate the initial positioning contour, which makes the initial contour located by ASM model evolutes to the direction of energy minimization. After the comprehensive use of the two types of models, experimental results show that the improved algorithm can not only strengthen the traditional ASM model labeled detection accuracy, but also can determine the contour lines between the points marked by the ASM model through the Snake model. It’s a better solution to face contour extraction of underground personnel issues.Meanwhile, on the basis of face shape modeling by the ASM model, this paper makes full use of the geometric constraints between the facial features of underground personnel, and obtains the standard shape and the required geometric transformation by the analysis of Procrustes; each instance data align with standard shape by the corresponding geometric transformation, and the face shape model is established by the shape generated from combining the local deformation and global change; the facial feature patch models are generated after training many times, the composite patch image that experiment needed is formed by combining each patch model; the face detection of underground personnel in the video is completed by modifying the related parameters. The experimental results show that: This paper has implemented the real-time and accuracy face dynamic detection of the underground personnel for noise pollution problems under the environment of a coal mine, and laid a solid foundation for the development of face recognition attendance system in the coal mine.