Optical system for evaluation of virtual firearms shooting training simulators, based on computer vision

The proposed system is an innovative solution that allows the implementation of shooting techniques training spaces using virtual reality and artificial vision concepts. The scenarios are emulated using development software called Unity. Silhouettes are virtualized images of typical targets useful to improve familiarity with the environment, precision and dexterity required by a shooter. The evaluation and qualification of the accuracy of the shot has as its main element a camera similar to a webcam, and under a total darkness environment, it is capable of evaluating with high precision the place of the silhouette where the infrared laser beam hit. The laser is adapted to a firearm through mechanical and electronic components. The image is processed using color recognition, edge detection, center of mass, and point detection; information that allows discriminating the contour of the laser and its center, to then assign a numerical value to the shot made. This ability to evaluate homographic images is another important contribution of this research. The results obtained in this investigation, after having made more than 10,000 shots, show the validity and significant contribution of the proposed system. The metrics presented are the product of actual shooting training exercises with people with little experience in the field.