提出了基于双相位编码联合变换相关器(JTC)的图像位移探测算法。随机产生一相位码对参考图像编码，与目标图像叠加作为输入图像，通过傅里叶变换得到联合功率谱，用同样的相位码对其编码并进行滤波，傅里叶逆变换后得到涉及位移矢量的互相关峰。给出了多帧图像间位移矢量的探测技术。相比传统的相关器探测技术，此算法能更有效地利用输入平面，将自相关项以噪声的形式散射在输出面，只保留一互相关项，有利于位移矢量的探测。比较了不同位移矢量探测技术，表明所提的技术更具有优势，均方根误差能够控制在0.1 pixel内。

An image displacement measurement technology based on double phase-encoding joint transform correlator (DPEJTC) is proposed. Firstly, a phase mask is generated electronically and applied to encode the reference image. The encoded reference image is overlaid with the target image to form the input image. Joint power spectrum (JPS) of the input image is obtained by a Fourier transform. Secondly, a filter is applied to JPS which is encoded by the same phase mask. After an inverse Fourier transform, a brightest peak in relation to the displacement between the reference and target image appears in the correlation plane. Displacement measurement of multiple images with this method is given. In comparison with the methods based on traditional joint transform correlator (JTC), this method can use the space of the input plane efficiently, disperse the auto-correlation item into system noises and remain the cross-correlation item only, which is convenient for detection. Comparing with different displacement measurement technologies, results show that our method takes edge over the other algorithms, and its root mean square errors (RMSE) can remain within 0.1 pixel.