Implementation of Single Image De-hazing System on DSP TMS320C6748 Processor

Abstract

Due to haze, the light reflecting from an outdoor object gets scattered before reaching the camera in almost all the outdoor scenes. They also decrease the photographic quality in the aerial atmosphere. It results in a massive delay in airlines and trains during the winter season. Hence, the hardware implementation for the removal of haze is an essential task in the field of computer vision. In this paper, we implemented a novel method for removing the fog from a single image using the dark channel prior method (DCM) with pixel by pixel operation. The DCM is adopted based on statistical data of the outdoor images which are haze-free. Based on the information from the DCM on hazy images, the thickness of haze is estimated. Thereby, a haze-free of a high-quality image is recovered. The implementation is done for single image on TMS320c6748 from Texas Instruments DSP kit and is simulated using Code Composer Studio v6.0. The validity of implemented techniques verified and compared with simulation during experimental approbation on the real hardware.