Research On Tracking Method Of Moving Target In Mine Complex Environment

As a major coal mining country,the safe production of coal mining enterprises is of great significance to China's social and economic development.The development and improvement of mine camera equipment and underground information transmission have laid the foundation for the realization of intelligent video surveillance and mine personnel tracking and positioning.However,due to the complex environmental background of the coal mine,the quality of the acquired mine image is relatively poor,and the resolution of the target is low.The application effect of the existing target tracking method is not ideal.Based on this,in order to further enrich the technology of intelligent processing of video images in underground mines,based on the synthesis of previous studies,this paper separately conducted research on mine image enhancement and target tracking technology.Aiming at the characteristics of mine images and the defects of existing algorithms,the improvement ideas are proposed.The specific work is as follows:(1)In terms of image enhancement,in order to improve the imaging quality of the mine image and enhance the observation effect,an improved Retinex mine image enhancement algorithm is proposed for the problems of color distortion,halo blurring,and over-enhancement when the traditional Retinex algorithm is used to process mine image.Firstly,the image to-be-processed was transform from RGB space to HSV space.Based on Retinex theory,the improved adaptive fast-guided filtering of V component is used to estimate the illumination,and then the reflection component can be obtained.A "S-shaped" function is proposed to perform illuminance equalization against luminosity components.Non-linear stretching is used for the reflection component to achieve details enhancement.Finally,the processed illuminance component and the reflection component are merged and converted back to the RGB space to obtain the final enhanced image.The algorithm is applied to the non-uniform illumination environment of the mine,and three representative algorithms are selected for comparison.The experimental results show that the proposed algorithm is superior to other algorithms in subjective and objective evaluation.It is concluded that the algorithm is superior in color,detail,and edge preservation,and can avoid over-enhancement and achieve effective enhancement of mine images.(2)In terms of target tracking,a mine target tracking method combining state estimation and improved Camshift is proposed.Firstly,the background difference is used to automatically obtain the initial target area to solve the problem that the target needs to be manually calibrated and can only track a single target in the original algorithm.According to the characteristics of the target to be tracked,the color and HOG features are combined to establish the target model to avoid the tracking failure problem caused by the interference of the background or similar targets.Finally,considering that the interlace occlusion phenomenon is easy to occur between the targets,Kalman filtering is used in combination to estimate the motion state of the target.At the same time,the estimated value of Kalman filtering is modified by the Camshift algorithm to improve the tracking accuracy.Through experimental comparison,it is verified that this method can adapt to the interference of the complex environment of the mine.And on the premise of ensuring real time,it can track multiple moving targets more accurately.The result analysis shows that the performance of this method is superior to other comparison methods.