Monocular Vision-Based Obstacle Detection Technique using Projected Grid Deformation

In this research, a novel technique for obstacle detection has been proposed. Typical obstacle avoidance and autonomous system applications make use of large number of different expensive active sensing equipment (LIDAR, RADAR, laser scanners and other sensors). While, this paper attempts to tackle the issue of obstacle detection using a relatively cost-effective set of active and passive sensors, namely a projector and monocular vision system. The proposed method makes use of grid projection on the surface for detection of stationary and moving obstacles on road with the help of assessing the deformation in grid line transformation. A host of different computer vision techniques have been utilized for grid line segmentation and detection to successfully detect obstacles and its distance from a fixed reference point. This paper examines the results of the grid projection on the object detection using different obstacle configurations to test the effectiveness of the proposed obstacle detection system.