Research On Matching For Degraded Images And Its Application In Video Extensometer

Image matching is an important research topics in computer vision and image processing technology. With the development of computer technology and artficial intelligence, digital image matching technology as part of the artificial intelligence, has been widely used in medical image processing recognition, face and fingerprint recognition. Video extensometer is an equipment for measuring the materials tensile properties using image matching technology.The principle of video extensometer is tracking the same point in un-deformed and deformed image, and then yields the displacements of the data point. However during image acquisition and transmission, the environment,imaging equipment and movement between object and camera will cause iamge blur. The image quality directly works on the following processing steps. In extreme cases, blur iamge will lose its valueBecause the blured image make it difficult for target matching, there are two types of solutions:Based on imae priori knowledge to restore image, and then matching the data point in the restored image. On the other way, we can extracted blur insensitive feature from the image, then based on these features to design matching algorithm.Based on the above ideas, this chapter proposes an image reconstruction algorithm based on the dictionary sparse representation, then matching on the reconstruction image. And introduce an matching algorithm based on the Local Phase Quantizations features to reduce the interference of the motion bring to the matching.At the end of paper, we introduce the video extensometer and its measuring system. And we apply both methods to video extensometer, the performance on high elongation material uniaxial tensile experiment also verifies the accuracy and robustness of the method...