Development Of Stereo Vision Sensor And Distance Estimation For Computer Vision Application (DE-CV)

Shamsul Fakhar, Abd Gani (2013) Development Of Stereo Vision Sensor And Distance Estimation For Computer Vision Application (DE-CV). Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Development Of Stereo Vision Sensor And Distance Estimation For Computer Vision Application (DE-CV) 24 Pages.pdf - Submitted Version Download (716kB)

Abstract

Stereo vision system is a practical method for depth gathering of objects and features in an environment. The purpose of this project is to design and develop a tool for stereo vision system with horizontal configuration and can estimate the range of obstacles or objects detected by stereo vision sensor. The application in this project is using autonomous guided vehicle as a platform to get the result of stereo vision system capabilities. The objectives are to optimize and to test the algorithm of disparity mapping. The construction of the software part and hardware part are run simultaneously. The integration and communication between them is using universal serial bus USB port. The sensor is using two Complementary metal–oxide–semiconductor as a stereo camera system that will be captured the images (left and right image). Because of lens distortion and camera displacements the next step is to rectify the images. Thus both cameras need to be calibrated first to get the camera parameters. The camera calibration is using Tsai’s method in Matlab. The results of camera calibration are used in rectification of stereo images. The stereo system calculates the disparity values by firstly, grabbing the images. Then, these images generate certain disparity values after entering the stereo matching process using Sum of Absolute Difference SAD algorithm in image processing software Matlab. The task of the algorithm is to build and to implement a stereo system for generating disparity mapping or depth maps. This project produces the less displacement of correspondence pixels shows that the pixels object is far away. To realize its calculation algorithms, this project needs a powerful processor and a motor controller module as well as CMOS cameras. At the end of this project, the AGV is capable to avoid the obstacles using the disparity mapping with an effective range about 0.3 meter until 4.5 meter.

Actions (login required)

View Item