|Purpose||The main purpose of this research is to develop the rectification algorithm of two images from separated cameras for ADAS application. Several temporal and geometric mismatches are dealt with in order to overcome the limitation of the conventional system which adopts an integrated camera module to keep the alignment of both images.|
|Contents||The scopes of this research are as follows;
• Rectification algorithm for temporal mismatch in stereoscopic image: The effect of mismatch in synchronization between two images are studied and the technique which estimates the time gap is considered. After that, the rectification algorithm for temporal mismatch is discussed. And the various road condition (straight, curved, up-hill or down hill) are considered.
• Rectification algorithm for geometric mismatch in stereoscopic image: The effect of geometric mismatch, such as orientations and directions in two images are examined and the rectification algorithm which compensates the spatial mismatch is developed. Also the differences in resolution, contrast, and white balance of two cameras are considered.
• Fast Algorithm for real-time application : The above both algorithms are integrated and optimized in order to be applied in real-time environment.
|Expected Contribution||It is expected to retain key technologies related to rectification of images with different spatial and temporal characteristics as well as camera parameters. They can be practically used in vehicle manufacturing and black box industry, since the limitation of the conventional integrated stereo camera module can be significantly removed. Also, this research can expand the use of stereo vision toward new area where multiple camera are avaliable.|
|2||스테레오 카메라 조정의 정확도에 미치는 체스보드 크기 및 위치의 영향||대한전자공학회 하계학술대회||2014-06-25||Phase 2
|1||소실점 및 소실선을 이용한 분리형 스테레오 카메라 정합 방법||대한전자공학회 하계학술대회||2014-06-25||Phase 2