3D Image Mosaic And Display Based On Light-Field Imaging

In the study of light field camera chip algorithm, based on light field imaging of three-dimensional image reconstruction technology research is a top priority. In the three-dimensional reconstruction of the process, although the mosaic effect can barely meet the demand by using of the existing splicing methods in image stitching stage. The total cost unsatisfactory as the time required is too long, which is applied to the chip for smart phones products. Therefore, to find a better and real-time high of stitching algorithms to improve the performance efficiency of products has become very urgent.SIFT (Scale Invariant Feature Transform) feature is the classical algorithm in the panoramic image stitching. Its advantage is that the image matching has good effect, but it has the long time-consuming problem. Facing this problem, A novel fast image stitching algorithm based on ORB (Oriented FAST and Rotated BRIEF) features is proposed in this paper. The algorithm firstly selects ORB algorithm for image feature extracting and matching to reduce the complexity of the feature extraction and matching and increase the speed of the image matching. RASANC (Random Sample Consensus) algorithm is then introduced to eliminate false matching points and improve the image matching effects. In the last weighted average method is used to speed up the image fusion. Experimental results show that our algorithm can keep almost the same mosaic image quality with the image stitching algorithm based on Sift features and Surf features, but our algorithm’s stitching time can reduce about74%and73%respectively with them.