A Research On Shooting Training System Based On HSV Color Space

In the conventional shooting training,the shortcomings of the traditional manual target reporting method gradually emerged.The manual reporting method is susceptible to the influence of weather,site and surrounding environment.Its reporting efficiency is low.The method is gradually abandoned due to high labor costs and hidden safety hazards.With the development of science and technology,technical researchers continue to make breakthroughs and innovations in the field of shooting training.The target reporting system is continuously improved through the use of computer technology and hardware equipment to overcome the shortcomings of manual target reporting.This paper proposes a shooting training system with target surface recognition and bullet hole detection.The system is implemented by combining hardware equipment and computer digital image processing technology.The shooting training system in this paper is mainly composed of three parts: image acquisition equipment(industrial camera),computer server,and mobile phone client.The three parts are connected in local area network.The target surface image from image acquisition device is transmitted to the server;the target surface and bullet holes are determined in the server by image processing algorithms;the recognition result is transmitted and represented.Users can observe the target surface and bullet hole detection result on one terminal device for each target channel.Thus,the training efficiency is improved.In this paper,a target recognition algorithm based on HSV color space and threshold segmentation is proposed.By separating the three channels(hue,saturation and value),the value of target hue is extracted to determine the target color threshold;the OTSU threshold segmentation algorithm is used to segment the saturation and the vale channel of gray scale images,respectively.The three-channel images that have been processed are finally combined to identify the target surface accurately.In this paper,the bullet hole detection algorithm is improved.The OTSU algorithm is used to segment the bullet hole from gray scale image.Bullet hole area is calculated to identify isolated bullet holes and adjacent bullet holes.In order to deal with the problem of target surface sway,a target surface similarity test combining SSIM structure similarity and the number of bullet holes is proposed to determine whether three consecutive frames of the target image are consistent.The experimental results show that the target surface recognition and bullet hole detection algorithm proposed in this paper based on HSV color space has good results under general conditions.For moving target,the algorithm decides the existence of moving target at the expecting target location,to shorten the response time.In the aspect of bullet hole detection,this paper analyzes the bullet hole detection algorithm and proposes the concept of bullet hole detection rate to prove the algorithm.The algorithm has high bullet hole detection accuracy under different lighting conditions.The adaptive threshold segmentation proposed in the improved bullet hole detection algorithm can effectively improve the bullet hole detection rate under different lighting conditions to overcome the deficiency of fixed threshold.The algorithm described in this paper can detect isolated bullet holes and adjacent bullet holes at the same time,which shortens the algorithm execution time.Statistics on the execution time of the algorithm demonstrates that the algorithm meets the real-time requirements.