An automatic recognition algorithm for bullet holes

An algorithm to extract the bullet holes from a chest bitmap out of doors real-time was proposed.A gray projection algorithm was used to extract the chest bitmap from the target bitmap.The mathematical morphology was used to stress the character of bullet holes and to overcome the disturbance of figures.Using mathematical morphology could resolve the overlapping of the bullet holes,then the fuzzy theory was used to extract the target holes accurately.For recognising the bullet holes,the histogram of the target bitmap was used to gain a threshold to exclude the pixels which could not become bullet holes to reduce the complication of computation and a strategy judging the distribution of the pixels around the target holes was used to exclude the invalidation bullet holes.Experiment results show that the algorithm can extract bullet holes accurately and adapt to various circumstances.If the size of the bitmap is 648×480,it only needs 150 ms to complete this algorithm.The new algorithm can also overcome the overlapping of the target holes.