Adaptive Headlight Beam Control Using Driver Gaze Saliency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing headlight control devices struggle to accurately control light distribution based on the driver's gaze, particularly when the object is difficult to detect or not detected by conventional cameras, leading to potential inadequacies in illuminating the intended area.
Innovation Solution
A headlight control device that generates a saliency map to determine the driver's gaze region and adjusts light distribution patterns based on the scene, using a saliency map generating unit, scene determining unit, and light distribution control determining unit to ensure the headlight irradiates the intended object with light.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection based on captured images is used to control headlight light distribution, then the headlight can be controlled to illuminate detected objects, but objects that are difficult to detect or not detected cannot be properly illuminated
Solution Approach 1:
The patent introduces an intermediary processing step between image capture and light distribution control. Instead of directly using detected objects to control headlights, the system first generates a saliency map that highlights regions of visual interest based on human gaze patterns. This intermediary representation allows the system to infer driver attention even when direct object detection fails, thereby improving the reliability of light distribution control.
2Productivity
If conventional object detection methods are used, then detected objects can be illuminated, but the system fails to illuminate objects in the driver's actual gaze region when detection fails
Solution Approach 1:
The patent inverts the conventional approach by not directly using object detection results to control headlights. Instead, it generates a saliency map from the captured image that represents regions of visual interest, and then uses this saliency map to determine light distribution. This inversion allows the system to capture driver gaze information that would be lost in conventional detection-based approaches.
3Adaptability or versatility
If headlight control is based on detected objects only, then illuminated areas match detected objects, but the driver's actual visual attention and needs may not be met
Solution Approach 1:
The patent changes the parameter used for headlight control from direct object detection coordinates to saliency map-based gaze region coordinates. By transforming the control parameter from detected object positions to regions of visual interest derived from human gaze patterns, the system achieves better adaptability to driver needs while maintaining reasonable accuracy in complex driving scenarios.
Data Source
AI summary
A headlight control device includes a saliency map generating unit to generate a saliency map indicating a gaze region that is a region easily gazed by a driver of a vehicle on the basis of an imaging signal indicating ahead of the vehicle, a scene determining unit to determine a scene that is a situation in which the vehicle is traveling on the basis of the saliency map generated by the saliency map generating unit, and a light distribution control determining unit to determine a light distribution pattern of the headlight on the basis of the scene determined by the scene determining unit, and output a light distribution control signal for controlling the headlight to emit light in the determined light distribution pattern.