With the increasing popularity of shooting sports, traditional manual readings have defects such as missed reports, false reports, and low security. To solve the existing problems, this paper utilizes deep learning technology to propose a dual-channel fusion Siamese network and detection algorithm for solving the difficult problems of microscopic hole position recognition and ring numbers reading on the target surfaces. By utilizing adaptive feature learning, combined with the closed area algorithm and ring value matrix, the calibration error in traditional methods is solved, achieving precis hole positioning and accurate ring number reading. The experimental findings indicate the high efficiency and accuracy of the above methods in target image processing and hole detection.