Detecting collisions in an unstructured environment through path anticipation

Detecting and avoiding a potential collision is the most important aspect when we talk about mobile objects of all types, ranging from autonomous vehicles, ship and aircraft navigation to robot manipulators etc. Image processing techniques could be used to provide solutions in such scenarios. Detection of a possible collision requires accurate information about the path of a moving object. The projected paths of all mobile objects in vicinity can illustrate if a possible collision scenario exists and its level of seriousness. Based on which, collision avoidance actions can be initiated. This paper presents use of image processing techniques for collision detection and avoidance scenarios using the statistical measurements