Trailer Angle Detection Using Pivot-Region Line Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle camera systems struggle to accurately determine the trailer angle with respect to the tractor during turns, especially when reversing, due to unreliable kinematic models and challenges in sensor detection in complex environments.
Innovation Solution
A method and system that utilize image processing techniques to obtain a birds-eye view from multiple rearward-facing cameras, identify lines intersecting a predefined pivot region between the tractor and trailer, and apply radial filtering to determine the trailer angle, enhancing accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If kinematic models are used to estimate trailer angle during forward travel, then the trailer angle can be determined without additional sensors, but the estimation becomes unreliable when the vehicle is reversing due to error accumulation
Solution Approach 1:
The system changes the measurement parameters based on operational conditions. During forward travel, it uses kinematic parameters (steering angle, speed, wheelbase) for angle estimation. During reversing, it switches to using visual parameters from camera images to detect trailer angle, avoiding the accumulation errors that occur with continuous kinematic integration during reverse operations.
2Measurement precision
If mechanical or limit-switch sensors are used to detect trailer angle, then direct measurement is possible, but the physical environment is not well suited for such sensors making detection difficult
Solution Approach 1:
The patent replaces mechanical sensors (limit switches, encoders) with an optical system using cameras and image processing. The trailer angle is determined by detecting lines in the image that intersect the pivot region and calculating their angular orientation, eliminating the need for mechanical contact sensors that are difficult to implement in the harsh physical environment of a trailer coupling point.
Solution Approach 2:
Instead of directly measuring the physical angle with a sensor at the pivot point, the system creates a visual copy of the trailer side in the camera image. It then processes this visual information by detecting lines and their intersections with the pivot region to infer the trailer angle, effectively copying the geometric information from the physical world into the image domain for measurement.
3Quantity of substance
If all detected lines in the image are used to determine trailer angle, then more data is available for calculation, but lines not intersecting the pivot region introduce errors and reduce accuracy
Solution Approach 1:
The system applies local quality filtering by giving different weights to different lines based on their spatial relationship to the pivot region. Lines that intersect the pivot region are assigned higher weight because they directly indicate the trailer's angular orientation at the coupling point, while lines that do not intersect are given lower or zero weight, as they may represent other features unrelated to the trailer angle.
Solution Approach 2:
The set of detected lines is segmented into two categories: lines intersecting the pivot region and lines not intersecting it. This segmentation allows the system to process and weight the lines differently, ensuring that only the relevant lines (those passing through or near the pivot point) contribute significantly to the angle calculation, thereby improving measurement precision.
Data Source
AI summary
A method of determining a trailer angle includes obtaining a birds-eye view image of a commercial vehicle based on image data from at least one camera mounted to the commercial vehicle that depicts at least one side of the trailer. The commercial vehicle includes a tractor and a trailer. The method also includes determining which of a plurality of lines in the birds-eye view image intersect a predefined pivot region which includes and surrounds a pivot point between the trailer and the tractor, determining a trailer angle of the trailer with respect to the tractor based on at least a first portion of the plurality of lines that intersect the predefined pivot region, and panning a video feed of images from the at least one camera based on the trailer angle. The determining the trailer angle includes excluding a second portion of the plurality of lines that do not intersect the predefined pivot region from the determination of the trailer angle, or weighting the first portion of the plurality of lines greater than the second portion of the plurality of lines in the determination of the trailer angle. A system for a commercial vehicle is also disclosed.