The weighing of heavy trucks is used in many industries. To obtain an accurate weight of a weighing vehicle, the vehicle needs to be accurately parked and does not exceed the boundary of the truck scale. For this, a vision-based parking line crossing detection system is proposed in this article. The proposed system uses two or more cameras placed around the truck scale to monitor if the parking vehicle presses the boundary of the truck scale. After the user defines the parking area in the interface, the system automatically generates an electronic fence of the area for the decision boundary. The system then uses the Mixture of Gaussians (MOG2) background subtraction algorithm to detect the foreground. Thus, the decision boundary is analyzed to determine if the vehicle is parked over the line. To verify the performance and robustness of the proposed algorithm, the system was evaluated under sunny days, nighttime, and rainy days. The proposed system was effective in detecting vehicle parking line crossing in different lights. In addition, the precision of the system reached 97.7, 100, and 96.6% on sunny days, nighttime, and rainy days, respectively. The system can be easily deployed in edge devices and can effectively enhance the efficiency and correctness of the weighing process for heavy trucks.
Visual-based Real-time Truck Scales Parking Boundary Crossing Detection
2024-04-19
768142 byte
Conference paper
Electronic Resource
English
Engineering Index Backfile | 1929
|