A Bullet holes detection algorithm based on wavelet transform and image fusion

A bullet holes object detection based on wavelet transform using fusion rules is proposed to overcome the limitation of the bullet holes detection using image difference technology that exists over-detection or leak-detection because of the noise interference and has harsh request for imaging condition. In the wavelet domain, it applied different fusion rules to the low-frequency and the high-frequency sub-image to revise the wavelet transform coefficients in order to increase the object coefficients while decrease or reset the non-object coefficients. Then, it used the maximum points to the low-frequency coefficient and reduced the noise of high-frequency coefficients, the final bullet holes images were obtained through wavelet composition. Experimental results show the proposed method can detect bullet holes object not only remove the noise effect but also enhance the edge, which improves the veracity and precision of the bullet holes detection.