Use of optimized computer vision algorithms are vital to meet real-time requirements of popular mobile platform like Beagleboard which use OMAP3530 .OpenCV is one of the extensively used computer vision library today. This library is optimized to take advantage of Intel architecture. However, when working on mobile platform like OMAP35x which also house heterogeneous DSP core C64x+, it is desired that this library give better performance using the on-chip DSP core. This project aims no narrow down this gap by porting low-level OpenCV API over DSP core. The execution of this project will be carried out in two phase. In phase one, some of the low-level OpenCV API like cvDFT(), cvSobel(), cvAvgSdv()and cvIntegral will be ported to DSP. In phase two, an application that make use of these accelerated libraries will be built to demonstrate the success of the project.