In order to solve the problems of capturing weak-texture features on the surface of large components and precision registration of multiple measurements, a compound measurement system integrating structured light and photometric stereo vision is adopted. The point cloud data of the overall shape of the workpiece surface are obtained by structured light measurement, and the normal vector information of the fine and weak texture of the surface is obtained by photometric stereo vision. On this basis, a new type of local feature descriptor which combines neighborhood point cloud coordinates and normal vector information is proposed, which can describe the surface features of weak-texture workpieces effectively and robustly. Extensive simulations and practical experiments verify the effectiveness of the proposed method, and its performance greatly surpasses the iterative nearest point algorithm based on traditional feature descriptors. The proposed method can effectively capture and describe the rich detail features of the weak-texture surfaces, construct robust and significant feature descriptors, and then greatly improve the matching accuracy of the measurement results and reduce the overall reconstruction error of large and complex components.