Distance Measurement using Stereo Vision

• Main Authors: Yu-ChiTsai, 蔡昱祺
• Other Authors: Chin E. Lin
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/81448392053891374920

碩士 === 國立成功大學 === 航空太空工程學系 === 102 === This thesis presents a distance measurement technique using computer stereo vision. The measurement result is quite accuracy within a defined range. The purposed approach sets up two cameras horizontally to capture the images, and adopts the similar triangles theorem to calculate the distance. In the measuring algorithm, the object in the left frame is selected manually to execute the matching algorithm to find the similar feature in right frame. It converts the disparity units into meter units and finally computes the distance of target. The proposed computer stereo vision is applied for human detection. The experiment module is established. A Histogram of Orientated Gradients (HOG) descriptor is used to distinguish the human features through the cascade classifier to determine the object for human beings. The proposed method in this thesis also develops the classifier for experiments. It accomplishes some outstanding contribution to target identification for collision avoidance to pedestrians.