A Multi-Image Stitching Method and Quality Evaluation
Xiaoshu Zhou, Haiyan Zhang, Yuejin Wang · 2017
The rapid growth of urban image scale and its resulting high cost of data storage have become the factors which will restrict the development of the smart urban security system. Image stitching technology can synthesize many smallview images into a picture with large view and high quality which helps to smart city overall planning and regulation. Also, image stitching removes the repetitive, redundant information to reduce the data amount in the image sequence and present the real world objectively and accurately. If two vectors existed in two planar planes cohere together, the relative position of two planar joining will be determined. Following to this principle, two-couple feature point matching method is proposed for urban scenery image stitching problems. According to perceptual characteristics of the Human Vision System (HVS), the image stitching quality evaluation criteria is set to verify the validity of the stitching method. The experimental results verify the rationality of the image evaluation index and the effectiveness of the mosaic method.