Image-based Distance Change Identification by Segment Correlation

摘要

Abstract Image–based distance identification is an interesting topic in image processing. It can be classified into three methods. The parallax approaches, the method with marker and the method with a priori knowledge of an object’s physi-cal size. The above methods are not easily implemented in an embedded system. This paper reports the development of an efficient image-based algorithm to iden-tify the change of distance between camera and the captured object by measuring the movement of objects in the image. The algorithm is designed to work on a single camera without the aid of image markers such as laser beams. The algorithm calculates the correlation of predefined segments in the image to detect object movements, and therefore identify the direction and magnitude of distance change. The algorithm was designed to be implemented in a quadcopter to identify its change of altitude. The performance of the algorithm was examined using a simulated environment, and is reported in this paper.