Automatic calibration method, device and equipment for external parameters of all-round view camera based on lane line detection and computer storage medium
A technology for lane line detection and automatic calibration, applied in computing, image data processing, instruments, etc., can solve problems such as affecting the labeling results, not suitable for real-time field operations, and cumbersome.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0090] The method for automatically calibrating external parameters of the surround-view camera based on lane line detection proposed in this embodiment needs to meet the following conditions:
[0091] 1. There is a map (world coordinate system) in the road section.
[0092] 2. It is necessary to provide an initial camera extrinsic parameter as an initial value for continuous updating. The translation matrix needs to be measured manually, and the error range of each direction should be within 10cm. An initial value must be provided for the rotation matrix.
[0093] 3. You need to drive steadily on this road section for a period of time, and at least one camera image in the surround-view camera needs to detect a lane line that meets the calibration conditions (the lane line is relatively straight and the detection effect is good).
[0094] 4. The collected data meets the images collected by the surround-view camera. Every pair at the same moment can be matched with features, an...
Embodiment 2
[0108] figure 2 is the change curve of loss, when the iteration reaches 100 times and the loss reaches 0.060, it converges. At this time, the lane line in the world coordinate system after the internal and external parameter transformation of the surround-view camera image is fitted with the corresponding map lane line. image 3 It is the fitting condition of the lane line in the world coordinate system and the corresponding map lane line after the internal and external parameters transformation of the surround-view camera image. The two white lines in the point cloud image are the lane lines on both sides of the world coordinate system. The two blue line segments are the lane lines on both sides in the world coordinate system after the surround-view camera image is transformed with internal and external parameters.
[0109]In this embodiment, the method of automatically obtaining external parameters for the first camera of the surround view camera uses the algorithm idea of...
Embodiment 3
[0123] This embodiment provides a device for automatically calibrating external parameters of a surround-view camera based on lane line detection, including a surround-view camera, a data screening module, a path solving module, a data checking module, and a feature matching module. The surround view camera is used to collect traffic scene data, and at least one initial camera external parameter is provided as an initial value for continuous update. The translation matrix is manually measured, and the error range of each direction is within 10cm. The rotation matrix provides an initial value; At least one camera image detects a lane line that meets the calibration conditions; the collected data meets the images collected by the surround-view camera, and every pair of features are matched at the same moment, and at least 8 matching feature points are matched; after the data is collected, first set the loss function , the loss function is based on the lane lines in the map corr...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


