Luminance compensation design using modified histogram equalization for driving assistance system under night vision

Night Vision Assistant can offer safety for the pedestrians at night, but the clarity of the capturing video has historically been an issue and needed to be improved. This paper focuses on applying MATLAB (Matrix Laboratory) for determining how the image should be modified through back-propagation neural network, and the image will be adjusted for better luminosity automatically through the modified algorithm of histogram equalization. The result shows that the modified image can be more easily discernable, as well as the distortion problems caused by overexposure and underexposure can be improved.