Vehicle Image Distance Estimation When the Lower Body Is Occluded
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle detection methods, such as shadow-based and learning-based approaches, fail to accurately measure inter-vehicle distance when the lower portion of the target vehicle is not visible in the driving image, particularly at close distances, which is critical for safe driving and autonomous driving systems.
Innovation Solution
A method that uses feature point tracking to detect and measure the inter-vehicle distance by acquiring driving images, detecting feature points in frames where the vehicle is not initially visible, and calculating the distance based on pixel position changes and image width ratios, even when the lower portion of the target vehicle is not photographed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If shadow-based or learning-based detection methods are used, then detection performance is improved when the lower portion of the vehicle is visible, but detection accuracy deteriorates when the lower portion is not visible in the driving image
Solution Approach 1:
The patent changes the detection parameters from relying on shadow patterns or full vehicle appearance to tracking specific feature points (corners, edges) that remain detectable even when the vehicle's lower portion is not visible. This parameter change enables the system to adapt to close-distance scenarios where traditional detection methods fail.
Solution Approach 2:
The patent introduces feature point tracking as an intermediary mechanism between image acquisition and distance measurement. Instead of directly detecting the entire vehicle, the system uses identifiable feature points on the vehicle as intermediaries to infer vehicle presence and calculate distance, enabling detection even when the vehicle's lower portion is obscured.
2Reliability
If the distance between vehicles becomes very close, then the lower portion of the front vehicle is not included in the driving image, but accurate distance measurement becomes critical for collision prevention
Solution Approach 1:
The patent extracts only the essential information needed for distance measurement - specific feature points on the vehicle - rather than requiring complete vehicle image information. By taking out only the necessary visual cues (feature points) and discarding the requirement for full vehicle visibility, the system can accurately measure distance even when the lower portion of the vehicle is not visible.
3Reliability
If feature point tracking is used instead of shadow-based detection, then detection reliability is improved in close-distance scenarios, but computational complexity increases
Solution Approach 1:
The patent segments the vehicle detection task into identifying and tracking specific feature points rather than processing the entire vehicle image or analyzing shadows. This segmentation reduces the computational burden by focusing only on key visual elements, thereby improving reliability in close-distance scenarios while managing computational complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for measuring an inter-vehicle distance is provided. The method includes acquiring a driving image photographed by a photographing device of a first vehicle which is being driven; detecting a second vehicle from the acquired driving image; detecting first feature points of a second vehicle region in a first frame corresponding to a frame in which the second frame is detected before a frame in which the second vehicle is not detected among a plurality of frames constituting the driving image, when the second vehicle is not detected from the driving image; detecting second feature points in a second frame corresponding to a current frame by tracking the detected first feature points; calculating a feature point change value between the first feature points and the second feature points; and calculating an inter-vehicle distance from the photographing device of the first vehicle to the second vehicle based on the calculated feature point change value.