Traffic Flow Determination Using Virtual Barriers

The patent introduces virtual barriers as intermediary elements in the monocular image space. These virtual barriers serve as reference objects to establish spatial relationships and enable speed calculation from 2D image data. The virtual barriers act as a mediator between the limited monocular imagery and the required 3D traffic flow information, allowing accurate speed determination without requiring multiple cameras or complex stereo vision systems.

The patent transforms the 2D monocular image space into a 3D traffic flow analysis framework by introducing virtual barriers with known real-world dimensions. This dimensionality transformation allows the system to infer depth and speed information from 2D images by comparing object sizes and positions relative to the virtual barriers, effectively adding a third dimension to the analysis without requiring additional physical sensors.

Virtual barriers are introduced as intermediary reference objects with known real-world dimensions. By comparing the apparent size and position of moving objects relative to these virtual barriers in the monocular image, the system can calculate speeds accurately. The virtual barriers serve as a mediator that bridges the gap between 2D image coordinates and 3D speed information.

The patent changes the parameter representation by introducing virtual barrier parameters (position, size, orientation) into the image processing framework. These parameters serve as reference values that enable the calculation of speed by comparing changes in object parameters over time. The system transforms speed determination from an impossible task in pure 2D images to a solvable problem by incorporating these additional reference parameters.

The processing means acts as an intermediary that integrates external data (map information, traffic rules, lane configurations) with sensor data in a unified framework. This intermediary processing layer manages the complexity by systematically combining different data sources and applying appropriate algorithms, making the integration process manageable and the results reliable without requiring complex reconfiguration of the entire system.

The system is designed with multi-functional capabilities that handle various types of data (sensor data, external data, map data) through a single processing framework. The processing means can perform multiple functions including object detection, speed calculation, traffic flow analysis, and rule-based validation, all within one unified system architecture. This universality reduces overall complexity by avoiding the need for separate specialized systems for each function.