A Blind Spot Early Warning System for Urban Traffic Passenger Vehicles
A technology for urban traffic and vehicles, applied in radio wave measurement systems, instruments, measurement devices, etc., can solve the problems of traffic hazards that are not identified and detected, lack the possibility of mass production, and have few identification types, so as to meet driving safety requirements. requirements, improve the accuracy of type judgment, improve the effect of robustness and accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0057] In this embodiment, the collision warning of urban traffic passenger vehicles is described in detail.
[0058] First, the target detection unit performs target detection on the image returned by the camera and the data returned by the radar at the same time. For the image data, the trained deep neural network is used to return the recognition results and the coordinate information of the target for various types of vehicles (including motorized and non-motorized vehicles) appearing in the image (the coordinate information includes but is not limited to the entire Outer contour point set, 2D frame information, 3D frame information of the car body, outer contour point set and 2D frame information of a single wheel). For radar data, the coordinate point and relative velocity of the moving target in the world coordinate system can be obtained through the preprocessing algorithm (including empty target removal, invalid target removal and stationary target judgment).
[0059...
Embodiment 2
[0076] In this embodiment, the pedestrian collision warning of the urban traffic passenger vehicle is described in detail.
[0077] First, the target detection unit performs target detection on the image returned by the camera and the data returned by the radar at the same time. For the image data, the trained deep neural network is used to identify all the human bodies appearing in the image. The special point of this patent is that in addition to the overall identification and detection of pedestrians, it also strengthens the individual identification of human body parts. The deep neural network of human semantic segmentation is used to complete the process, because the pedestrian collision event is most likely to occur in the process of stopping the bus at the stop. Pedestrians appearing close to the vehicle body (with each other obscured or only partially visible) can be recognized.
[0078] The feature of the present invention for pedestrian recognition lies in the use o...
Embodiment 3
[0096] This embodiment further describes the lane departure warning. In the lane departure warning, the target detection unit uses the deep neural network to detect the contour points outside the lane line (including the left and right lane lines in the own vehicle lane and the lane lines of the left and right lanes adjacent to the own vehicle lane). In the fusion calculation unit, after the point set is transformed into a unified world coordinate system by projection, the curve equations of all lane lines can be obtained by fitting the quadratic curve, which can be solved simultaneously with the straight line equation by fitting the forward direction of the vehicle. The coordinates of the collision point, the relative distance from the coordinates to the coordinates of the touch point of the four tires of the vehicle can be calculated. In the alarm unit, the distance is divided by the current vehicle speed to calculate the line pressing time T, and the minimum value Tmin is s...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


