Papers by Author: Nai Guang Lu

Abstract: A binocular point matching method using affine transformation is presented in this paper to deal with matching ambiguities. The epipolar geometry is used to determine all the possible matching pairs to form an initial correspondence data set. Then, an affine registration model with four parameters that is invariant to scaling, rotation and translation is built using the Random Sample Consensus (RANSAC) method to describe the coordinate transformation between the two members of correspondences. Finally, correspondences are picked out using minimal nearest neighbor distances based on the geometric similarity between the right image points and the transformed left ones. The proposed method is applied to measure the profile of a 3.5m parabolic reflector of an inflatable antenna and proved to be able to handle the extra or missing point matching problems coursed by occlusion and sheltering. Satisfactory results are obtained with high correct rate for matching most image pairs despite the significantly different viewpoints, which indicates its validity and application feasibility in automatic stereovision measurement field.

Abstract: The inflatable space antenna is gradually used in various spacecrafts because it is portable and foldaway. It is usually made of thin-film materials and has a flexible surface, so that measuring force is not tolerable in measurement process. Close-range photogrammetry is considered as an optimal solution because of its advantages of non-contact operation and fast data acquisition. To improve measuring precision, a method combining bundle adjustment algorithm and the distance constraint is presented in the paper. Two experiments under different conditions are accomplished and experiment results are compared. One experiment is completed with the distance constraint and another without. The experiment results are compared by two parameters: tightness and residuals. The measured object is an inflatable antenna with a 3.5-m diameter.

Abstract: Many vision tasks such as 3D measurement, scene reconstruction, object recognition, etc., rely on feature correspondence among images. This paper presents a point matching method for 3D surface measurement. The procedure of the method is as follows: (1) rectification for stereo image pairs; (2) computation of epipolar lines; (3) sequential matching in vertical direction; (4) sequential matching in horizontal direction. The fourth step is performed to deal with the ambiguity in dense areas where points have closer vertical coordinates. In the fourth step a threshold limit of vertical coordinate difference is designed to determine those points potential to cause ambiguity. This method was applied to the 3D surface measurement for an inflatable parabolic reflector with validity of point matching up to 100%. Experiment results show that this method is feasible in application of sparse point matching for continuous surface measurements.