PROCEEDINGS ARTICLE | January 31, 2013
Proc. SPIE. 8759, Eighth International Symposium on Precision Engineering Measurement and Instrumentation
KEYWORDS: Detection and tracking algorithms, Data modeling, Imaging systems, Sensors, Image processing, Error analysis, Image acquisition, Feature extraction, Optical alignment, Target acquisition
This paper persents an laser and target alignment sensor (LTAS) used to automaticlly position for the ICF experimental target. LTAS which is based on optical conjugate principle can locate the targets online with its four imaging measurement systems. For the requirement of automatic positioning, presents a new method which is called offline test and online match method. With the help of target offline test device, the reference image features are provided, online images and reference images are matched in real time based on images point features, SIFT descriptors are used for extracting target feature from image which is invariant features to image scale and rotation. According to image feature, we modify the original feature vectors with 128 dimensions to a feature vectors with 32 dimensions. A fast nearest neighbor algorithm is used for feature match, and the Best-Bin-First (BBF) algorithm is used to search nearest neighbors of points, and then RANSAC algorithm is used to remove the mismatch points, A least-squares solution is used to calculate parameters between the reference image and the online image, according to the matching parameters, the target is automatic adjusted to proper attitude. By matching experiment, the position matching accuracy is within 1 μm, the angle matching accuracy is within 0.1 degree, matching time cost is less than 1 second, and meets the real-time matching requirements.
