Implementation of barrel distortion correction on DSP in real time

Camera plays a vital role in intelligent embedded image processing systems. But unlike human eyes, it creates distortions in the image captured, most commonly called as lens distortion. The uneven magnification over the lens surface is responsible for lens distortion. Distortion conveys incorrect information about the coordinates in the image frame with respect to real world coordinates and thus decreases the positional accuracy of the processing in the complete system. In video or image processing applications like motion detection, satellite tracking, automatic vehicle parking, ADAS etc. the lens distortion needs to be eliminated. Hence, to improve accuracy of such applications, implementation of a barrel distortion correction algorithm in real time is presented in this work. The algorithm is implemented on the Analog Devices' BF561 Digital Signal Processor for real time distortion correction, and the results are discussed. The results shown include the outputs of algorithm on digital signal processor platform.