Optimization of image processing in video-based traffic monitoring

The video-based traffic monitoring systems have been widely used. The system usually reads real time monitoring video and converts them into images for processing. However, such systems are often limited by image processing algorithms and don't behavior as expected. We hereby study optimization approaches for different image processing stages. As getting a binary image is important and fundamental to image processing, we first present an adaptive thresholds approach for binary conversion. The approach chooses thresholds according to each pixel and its space information. Then we introduce a three-dimension Gaussian filter for denoising. We also propose a Gaussian model based background generation approach so that we can update background at any time, thus getting rid of a major limitation of background subtraction approach. Moving objects shadow detection and removing mechanism is also added in. We put forward a burst noise eliminating algorithm that uses several continuous frames to wipe off sudden noise in real world monitoring. Finally, we assert that removal of the reflection light band is to alleviate negative effect from high-energy reflection light. We conclude that adding in white light will reduce color saturation. In the area affected by reflection light band, there are much more white light involved than other area. Based on the conclusion, we use Gaussian model and an improved four direction Sobel operator to achieve reflection light band elimination from image.