Fuzzy Logic Based Vehicular Congestion Estimation Monitoring System Using Image Processing And KNN Classifier

摘要

Vehicle is one of the most valuable mode of transport human developed. This allows us to travel faster from point to point ordifferent multiple destinations. But through years, population increase and congestion occurs on public road. The study proposes a differentmethod of image processing – morphological feature extraction, KNN classifier and fuzzy logic on classification of common vehiculartransport mainly found on the road namely bus, cars and motorcycle. The images were taken using a smartphone camera 8mp and 12inches range from the 240 sample miniature vehicles. It is then processed using a laptop with MATLAB 2012 installed. The extractedfeature is area and shows ranges of 42,000 to 57,000 for busses, 13,000 to 35,000 for cars and 4,000 to 13,000 for motorcycles. The dataextracted were used for KNN classification for determining the vehicular type and for fuzzy logic decision making as the output is thedegree of congestion which is dependent on the road area of the image taken and decision is converted to percentage (0-40% light, 41-70% moderate, 70-100% heavy). Input parameter is the number of area on a certain image which is rated as few (0-40%), moderate (41-70%), heavy (71-100%) for all vehicle samples. Consequently, the result of the study shows a great potential on vehicular congestionmonitoring system using image processing, KNN and fuzzy logic algorithm used.