An SSD Algorithm Based on Vehicle Counting Method

Detecting and counting vehicles in transportation surveillance videos are essential for applications ranging from traffic queue detection to incident and vehicle identification. In this paper, we propose a vehicle counting method based on Single Shot MultiBox Detector (SSD). We perform vehicle detection, vehicle tracking and counting tasks via SSD. The SSD has the capacity to extract meaningful features of blob of vehicles, and predicting the location and classification of vehicles. Tracking moving vehicles can be achieved by comparing the blob of vehicles and measuring the minimal distance between two temporal pictures. Furthermore, given that the virtual coil method is easy to implement, we utilize this method to count the number of moving vehicles with fast speed. Finally, the experiment results are provided to illustrate that the effectiveness of the proposed method on counting vehicles with above 87% accuracy of real-time processing speed.