The detection algorithm research of random contrary color object Based on image processing

In the application of CCD optical target, we usually use the shadow image, which is formed by the projectile shading the background light, to achieve the capture of the projectile and the recognition. However, in the practical application, a number of projectile may be randomly mixed with tracer bullets and other projectiles trailing with light trail, making the linear array CCD imaging is no longer a simple shadow image, instead, forming invert color target image, which will produce great distraction for the recognition of projectile, even fail to identify the target signal and appear the omission phenomenon. According to the characteristics of the target image, a algorithm, which can detect the bright target and the dark target, can be proposed. The algorithm is that dropping shadow, firstly, image on the horizontal direction, then detecting the maximum point of local maximum of the de-noising projection signal to obtain the target line; detecting signal target break point in this line, and further to the center location of the projectile. Experiments show that the algorithm is reliable to identify and calculate the center of two images. It is an effective and practical processing algorithm.