Real-time enhancement using multi-linear adaptive gamma correction (MLAGC) for better night driving

In this paper, a real-time video-stream enhancement scheme is proposed along with a hardware prototype development. The proposed technique employs a multi-linear adaptive gamma correction method to locally enhance the dark video frames in presence of high intensity optical sources. The proposed algorithm employs three linear functions corresponding to low, medium and high intensity regions for proper enhancement of different intensity regions. The slope and range of the linear functions are automatically derived from intensity distribution and hence the overall multi-linear function becomes highly nonlinear and adaptive to varying constraints. Consequently, the input–output mapping function of the algorithm works as a spreading function in the low intensity range for contrast improvement and preserves the information of the high intensity regions to avoid information loss due to over-enhancement. The experimental results show that the proposed algorithm exhibits better performance than various existing algorithms in terms of subjective and objective measures.