Study Of Military Automatic Target-scoring System Based On Machine Vision Technology

The technology of machine vision detection is used for automatic indication ofshots in the training of military shooting,it can prevent the occurrence of trainaccidents, also can get the distribution of the point of impact during the gunnerytraining of soldiers,can remind us to take appropriate training means to improve thefiring level of soldiers.In this dissertation，the train of chest ring target shooting in field firing training isour object, an automatic indication of shots system is studied based on machine visiontechnology. The main research contents include the following aspects:1. Building the automatic hardware and software platform of target system.Combine the training characteristics of shooting subjects in troops and the designrequirements of the indication of shots system, we propose the design scheme ofautomatic indication of shots system based on the machine vision technology, andcomplete the design of related software and hardware systems.2. The target image preprocessing. Due to the silhouette paper image which istaken in by the CCD camera include the target surface area,the background regionand the noisy stream, we need to get a smooth image of the target surface area byusing the image preprocessing technology. Firstly, to detect the specific color featureedge of the silhouette paper image, then to get the target surface contour by using theHough calculation, to complete the image locating of the target surface. To filtertarget surface color images by using morphological filtering operator, not only we canget the smooth target surface image,but also retain the color information of the targetsurface image.3. The recognition and extraction of bullet hole region. In order to solve theidentification problem of bullet hole of chest ring target paper image quickly andaccurately, this dissertation proposes a identification method of bullet hole of chestring target by combining bullet hole edge character and color characteristics. Firstly,to detect the image edge of chest ring targe by using the morphological edge detectionalgorithm, to enhance edge image, to improve the contrast between the edge imageand the background, and then to achieve the bullet identification of chest ring targetpaper image by extracting the colour characteristics of the data of bullet holes area.4. To determine the number of holes ring. In the process of determining the pointof impact rings,firstly, to use the least square method of curve fitting to detect thebull’s-eye of silhouette paper, and to use the line segment to connect the bull’s-eye of silhouette paper with elastic centre,then to determine the point of impact rings bythe way of estimating the pixel gray value step number in the line segment.In this dissertation, we design the software and hardware platform of automaticindication of shots system by combining the reality of the force firing training,anddiscuss the image preprocessing, the bullet identification and the determining thepoint of impact rings in the automatic target system thoroughly,and these algorithmscan be transplanted to automatic indication of shots of other target type by makingsome modifications.