面积-极坐标变换在光学相关目标识别中的应用

Application of Area-Polar Transform in Optical Correlation Target Recognition

应用联合变换相关器对目标进行探测和识别时,由于实际探测到的目标相对于参考图像存在比例、旋转、平移畸变,因此在传统的光学相关器中很难实现目标的探测和识别。为了更好地探测到目标,提出了面积-极坐标的算法,该算法利用相似图形之间的尺寸与面积存在满映射函数关系,实现了比例不变光学探测,并结合极坐标变换实现旋转不变光学探测,最后,对变换后的联合图像进行小波边缘提取,既能提取到清晰的图像边缘,又能有效地抑制噪声,不但提高了相关峰,并使得相关峰更为尖锐,成功实现了大尺度混和畸变目标的探测和识别。作为实例,对混合畸变目标进行了计算机模拟和光学实验,验证了该算法在大尺度混合畸变光学相关探测中的可行性。

Owing to the existence of the scale,rotation and translation distortions between detected object and reference object,it is difficult to detect and recognize the target object for classical joint transform correlator.To improve the performance of target detection and recognition,area-polar transform algorithm is proposed.Firstly,by employing the relationship of area mapping between similar objects,we can realize the scale invariant.Secondly,rotation invariant is implemented by polar transform algorithm.Finally,after performing area mapping transform and polar transform,the edge of joint image is detected by wavelet transform,which can be used to extract the edge features and also reduce image noise.The proposed algorithm not only improves effectively the correlation peak,but also realizes detection and recognition of large scale and rotation distortion.The experiments are performed and the results are in accord with the theoretical simulations.The experiment proves the validity of proposed method.