Abstract: There are certain hazardous areas under the mine that do not allow workers to enter without safety measures, however, due to some environmental or human factors, the commonly used warning signs are not strong enough, resulting in the occurrence of underground accidents.Therefore, an automatic detection method based on computer vision is used to detect whether workers enter dangerous areas. The method first uses a kernel correlation filter algorithm KCF (Kernel Correlation Filter) to track the target,then use the binocular vision principle to measure the tracked target.Simultaneously the Kalman filter algorithm is used to optimize the obtained distance information,which make the ranging method a certain robustness for the occlusion target, and ensure that the target is outside the dangerous area.The downhole target tracking and ranging method has the advantages of low price, low calculation amount, large measuring distance range and high accuracy of ranging.