Application Research On Machine Vision Based Automatic Target-Scoring System

The purpose of this paper is do some research in automatic target-scoring system based on machine vision, which is met with request of ball firing in troop. The main object of system is carried out automatic target-scoring in computer, reduced the effect of disadvantage factors in traditional training, such as low efficiency, low security, guaranteed training security, decreased the difficulty of safeguard, enhanced the efficiency of training. So it is significant for building modern troop to research the automatic target-scoring system.In the shooting, cameras are used to continuously observe on targets. According to the characters and change in collected images of target goal, using imagine matching, imagine abnormality modified, then the current frame compared with the front frame, got out the new ball point in image of target. The position of ball point can be got by coding in target area. In the process of observation, technologies such as gray-scale image matching, image enhancement are used to solve the efect on identification accuracy by interfere factors such as light change, wind etc. which can improve the performance and accuracy.Before digital computer image processing, it is necessary to carry out image acquisition. Image acquisition converses analog consecutive image into digital discrete images. After receiving the digital image, target digital image pre-processing improves the image quality, due to prepare for image recognition. The pre-processing technology is used in the system such as gray-scale image, histogram equalization, binary image, median filter and other filter etc. After pre-processing, by image matching and image recognition, the ball point is found in target goal, and then calculated the ball point's area. In the process, using multi-resolution matching algorithm, to solve the low efficiency of system identification in the situation of larger image distortion.In the system debugging, image processing algorithms should be tested respectively. The algorithm parameters was made, then integrated debugging, adjusted parameters, testified real-time character of system; in addition, other function modules are tested. Tests showed that algorithms are efficient, modules are reliable.