A Region of Interest Extraction Approach Based on Adaptive Threshold in Intelligent Transportation System

It is important to extract ROI (Region of Interest) based on vision image processing in intelligent transportation system. A simple and novel approach is proposed to extract the RIO in transportation scene images. Firstly, a simple method is used to preprocess the input images; Secondly, interest pixels extraction function is defined to detect all the interest pixels from scene image using adaptive threshold mechanism; after that, these interest pixels connective each other are grouped to interest regions, the non-interest pixels enclosed by interest region are filled as interest pixels; Lastly, the regions selecting using region area threshold are as the extraction result of RIO. The experiment result is presented in the last part of the paper and the proposed approach is feasible and valid.