Implementation of Vehicle Around View Monitoring (AVM) System

• Main Authors: Chao-An Jhong, 鄭朝安
• Other Authors: Hsien-Chou Liao
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/6f33z6

碩士 === 朝陽科技大學 === 資訊工程系 === 102 === Cameras are popularly used in the vehicles. Around view monitoring (AVM) system is one of the emergent applications. The around view of a vehicle can solve the blind spot problem. For the previous AVM systems, there exists a clear boundary between the images of two adjacent cameras. Therefore, an implementation of the AVM system is proposed to overcome the above problem. Firstly, the feature points of all the images are detected by using the Harris corner detection method. Then, the maximum correlation rule and RANdom SAmple Consensus (RANSAC) method is used to establish the correlation between two sets of feature points. A homography matrix is then established. The matrix is used to stitch two or more images together into a single, large image. Then, the single image is transformed into a bird-eye view. Various image resolutions are used in the experimental study, the maximum FPS (frame per second) is about 2.4 to 11.7. A real-time AVM system with 30 FPS can be realized in hardware in the future.