中文摘要：

针对传统质量图易受干涉条纹宽度影响,及其在真实相位差等于2kπ的陡峭区域的低识别能力,提出了一种依靠法向量一致性来度量相位质量的方法。在3×3的缠绕相位窗口中,以中心点为参考点,进行局部相位解缠。解缠后,中心点与每对窗口边界上的相邻点构成8个三角形,将8个三角形的单位法向量合成后的法向量作为局部解缠相位平面的法向量,8个三角形的法向量与合成法向量之间的夹角和作为中心点的相位质量。对仿真数据和真实InSAR干涉纹图的处理结果表明：在质量引导算法中,新质量图比传统质量图具有更精确的路径引导能力。

英文摘要：

A new quality map is introduced for wrapped phase image in this paper. In order to reduce the affluence of the slope effect in the conventional quality map and improving edge-detection ability in the steep discontinuity areas where the real phase dif- ference is 2 kπ, a quality map derived from the consistency of normal vectors is proposed. In 3×3 window of wrapped phase image, the center is selected as the reference point and then the local unwrapping is performed in the window. After local unwrapping, eight triangles are formed by the center and every pair of neighboring points on the border of the window, then the syntheti-cal normal vector formed by the eight normal vectors of the corresponding triangles is taken as the normal vector of the local unwrapped surface. Finally, the quality value of the center is defined by the sum of angles between the eight normal vectors and the synthetical normal vector. The tests performed on simulated data and real InSAR wrapped phase image show that the new quality map has the better guiding ability than the conventional quality map in the quality-guided phase unwrapping algorithm.