| Summary: | There has been an increasing trend in the utilization of Visual Data. This is due to its ability to overcome limitation of the naked human eyes. Thus, with the increase in interests to utilize Visual Data , it caused an increase in research project which involves the extraction of visual data from images.
Stereo Vision (SV) systems allow one to perform Image Acquisition and Image Reconstruction of a real-life object. SV system is a type of binocular vision system which aim to extract three-dimensional(3D) information from two dimensional(2D) images, overcoming the limitation of monocular vision system. With the 3D information of the object, it can serve as input control parameter for other machineries in the manufacturing plant.
The aim of this project is to improve the current software code of the existing SV System by using Python Programming and Open-Source Computer Vision Library (OpenCV). Experiments will be conducted, to test the code’s ability to calibrate the cameras of the SV System and reproject the 2D images into 3D point cloud of the real-life object. In conjunction to the software exploration, the project will investigate the mechanical design of the current SV System, with the aim to improve the design to allow ease of usage and more accurate parameter tuning. The idea is to address several disadvantages found on the current system during the operation of the SV System.
This report introduces Stereo Vision and Stereo Vision systems, the set up in the lab, formulas to select the components of the system, the software used in carrying out the experiments, results obtained from the experiments. Finally, the report will end off with the discussion of the results, suggestions to improve the mechanical design of the set-up, possible limitation faced during the project and suggestion for future works.
|
|---|