Research On Image Enhancement And Target Tracking In Underground Mine

Coal occupies an important strategic position in the development of our country,so the safe mining of coal mines is of great significance to our country’s social and economic development.With the rapid development of the computer industry,the use of smart devices to monitor underground mining and employee life safety has made great progress.However,due to the objective and complex environment of the mine,the development of technology mainly encounters the following two bottlenecks.The first is that the illumination of the images collected by the monitoring equipment is generally low and the interference is large.Secondly,the mine tunnel is relatively long and long,and the phenomenon of target loss is easy to occur in the target tracking process.The actual application effect of the existing target tracking method cannot be achieved.expected.On the basis of previous studies,this article will specifically propose the following solutions to these two problems.(1)Image enhancement.In order to make the target tracking under the mine more accurate,image preprocessing is extremely important.Coal mine safety monitoring technology is an important part of the coal mining process.Underground video monitoring provides an important guarantee for coal mine safety.Therefore,the quality of monitoring images directly determines the effectiveness of monitoring.The underground environment of the coal mine is harsh.Due to the darker light and more dust,the images acquired by the monitoring and acquisition equipment are prone to problems such as uneven illumination and poor contrast.This paper proposes an improved Retinex algorithm based on HSV spatial fusion,which not only achieves the purpose of image enhancement,but also reduces image distortion and blur.First,the preprocessed image is transformed from RGB space to HSV space,and the illuminance component is estimated by the Retinex principle using an improved bilateral filtering algorithm,the reflection component is nonlinearly processed,the saturation is corrected,and the color is true,and the processed component Perform fusion,and finally inversely convert the image to RGB space to complete the image enhancement process.Experimentally verified,this paper proposes an improved Retinex algorithm to be applied to the non-uniform illuminance environment in underground mines.It compares with the three typical algorithms to judge the image average,standard deviation,peak signal-to-noise ratio and information entropy..Experimental data proves that in the process of image enhancement,the algorithm proposed in this paper not only improves the brightness and contrast of the image,but also effectively suppresses the halo,noise and edge blur phenomenon of the image,and provides more sufficient safety for mine production.Guaranteed.(2)Target tracking.This article uses the main YOLO algorithm to track mine targets.The advantage of the YOLO algorithm lies in its accurate calculation and fast execution speed,which can meet real-time target tracking in the mine.Aiming at the shortcomings of the low accuracy of the YOLO algorithm under low illumination,this paper uses the pyramid model to fuse the YOLO algorithm and improves the loss function of the YOLO algorithm.After comparing the method in this paper with the Camshift and original YOLO algorithms,the experimental data shows that the method in this paper can meet target tracking in the complex environment of the mine.It not only improves the accuracy of the amount,but also guarantees the real-time premise.The result analysis shows that the method The performance is better than the comparison method.