D.M. Keathly, T.-I. Yang, and Y. Wan (USA)
A fusion-based approach to parking lot monitoring is presented using low-cost video cameras. The system monitors the state of each individual parking space on a continuous basis and provides reporting capability when triggered by an lot entry sensor system. A trio of image processing techniques, coupled with a unique 3D volumetric approach to deal with adjacent vehicle occlusion, is used to detect the presence of vehicles in the space. The individual detection results are fused in a voting mechanism to reach a final decision. The use of multiple techniques achieves a result that is less susceptible to changes in ambient lighting, shadows and other potential causes of misdetection. Results for the individual and fused processes are presented which show an overall benefit to be derived from the techniques described.
Important Links:
Go Back
