Research On Recongnition Technology With Occlyded Target

Optical correlation detection has advantages in speed and accuracy. In this paper, real-time joint transform correlator (JTC) is used to correlation recognition. After correlation processing, target and template joint image is generated correlation peak image. Through correlation peaks judgment, the target location is determined, the target recognition is realized. But in actual applications, the target is often occluded, resulting in the target information losing. Target recognition accuracy is affected. So it is necessary to take on the occluded target recognition research.Based on the characteristics of occluded target local information losing, in this paper, the wavelet theory is applied in joint image pre-processing. Combining multi-scale wavelet transform, multiwavelet transform with modulus maxim a method, different kinds of background occluded target edges are successfully extracted. After processing, compared with correlation peak contrast. background information effect is reduced. Then the occluded target recognition accuracy is improved. In this paper, traditional histogram equalization is also improved to enhance local information to realize occluded image local processing, in this way the target information is effectively highlighted.Many optical experiments show that wavelet analysis and local histogram equalization processing is applied in occcluded target recognition, and the correlator identification accuracy is improved, realizing the occluded target recognition.