Distracted Driving Detection via Gaze-Scene Target Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distracted driving detection technologies have low accuracy as they primarily focus on detecting whether a driver looks straight ahead or shows fatigue features, failing to accurately determine when a driver is 'minds-off' and thus pose a safety risk.
Innovation Solution
A method that involves receiving facial and external scene images, calculating the driver's gaze point coordinates, and comparing them to scene targets using computer vision and inertial measurement unit data to determine if the driver is distracted, converting subjective distracted driving behavior into quantifiable information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing technical solutions focus on detecting whether a driver looks straight ahead or shows fatigue features, then the detection process is simple, but the detection accuracy is low and cannot accurately detect minds-off states
Solution Approach 1:
The patent combines multiple detection technologies including facial image recognition, eye tracking, head pose estimation, and scene target detection into a unified distracted driving detection system. By merging these separate detection functions, the system achieves comprehensive monitoring of driver attention state, enabling accurate detection of minds-off conditions that single-method approaches cannot detect
Solution Approach 2:
The patent introduces an intermediary processing layer that correlates driver gaze coordinates with scene target coordinates. This intermediary step transforms raw detection data into meaningful attention assessment by determining whether the driver's line of sight intersects with relevant scene objects, thereby accurately identifying distracted states without requiring complex direct observation
2Reliability
If the system only detects fatigue features like yawning and blinking, then the implementation is straightforward, but it cannot detect distracted driving before fatigue sets in
Solution Approach 1:
The patent implements preliminary detection of distracted driving behaviors by monitoring driver gaze direction and scene target attention before fatigue symptoms manifest. By detecting whether the driver is paying attention to relevant scene objects (traffic lights, pedestrians, other vehicles) in advance, the system can issue early warnings and prevent safety incidents before fatigue sets in
Solution Approach 2:
The patent replaces traditional mechanical fatigue detection methods with computer vision-based gaze tracking and scene target recognition. This substitution enables non-contact, automated detection of driver attention state by analyzing facial images and calculating gaze coordinates, making the detection process more sophisticated yet easier to implement at scale
Data Source
Figure 1~3
Figure 4~5
AI summary
The disclosure relates to the technical field of safe driving, and provides a distracted driving detection method, a vehicle-mounted controller, and a computer storage medium. The method includes: A: receiving a facial image of a driver and an external scene image that are synchronously acquired; B: obtaining coordinates of a scene target in the external scene image by using an object detection algorithm, where the scene target is an object capable of providing risk perception for the driver; C: calculating coordinates of a gaze point of the driver in the external scene image based on the facial image of the driver and the external scene image; and D: determining, by comparing the coordinates of the gaze point with the coordinates of the scene target, whether the driver is in a distracted state. The vehicle-mounted controller includes: a memory; a processor; and a computer program stored on the memory and executable on the processor, where the computer program is executed to cause a distracted driving detection method to be performed.