Driving Video Fusion for Accurate Dangerous Behavior Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining driving behavior based solely on vehicle running status are one-sided and prone to misjudgment of dangerous driving behaviors, affecting the accuracy of insurance strategies.
Innovation Solution
A method and apparatus that utilize a driving video, including road and vehicle images, in conjunction with vehicle running information, to comprehensively determine dangerous driving behaviors using AI algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If driving behavior is determined based solely on vehicle running status, then the determination process is simple, but the accuracy of dangerous driving behavior identification deteriorates
Solution Approach 1:
The patent combines multiple data sources including driving videos, road images, vehicle images, and vehicle running information into a unified determination system. This merging of diverse information sources enables comprehensive analysis of driving behavior, allowing the system to accurately distinguish between intentional dangerous driving and normal driving situations like emergency braking in congested areas.
Solution Approach 2:
The patent introduces AI algorithms as an intermediary that processes and analyzes the collected driving videos and vehicle information. These algorithms serve as a mediator between raw data and behavior determination, automatically identifying dangerous driving patterns while filtering out normal variations, thereby improving accuracy without requiring manual intervention.
2Device complexity
If only vehicle running status is used for determination, then data collection is simple, but misjudgment of dangerous driving behavior occurs
Solution Approach 1:
The system merges multiple data collection channels including video cameras for road and vehicle imaging, sensors for running status monitoring, and AI processing systems. This multi-source data collection approach ensures comprehensive and reliable information gathering, reducing misjudgments in insurance strategy development.
Solution Approach 2:
The patent implements a feedback mechanism where the determination results are continuously refined based on analyzed data patterns. The system learns from accumulated driving behavior data to improve its accuracy over time, ensuring that insurance strategies are based on progressively more reliable assessments of actual driving risks.
Data Source
Figure 1~2
Figure 3A
Figure 3B
AI summary
Embodiments of the present disclosure disclose a driving behavior determining method, apparatus, and device, and a storage medium, and belong to the field of Internet of Vehicles. The method includes: obtaining a driving video, the driving video including a road image and vehicle images of front and rear vehicles; obtaining vehicle running information, the vehicle running information being used to represent a running status of a vehicle; and determining whether a dangerous driving behavior exits according to the driving video and the vehicle running information at a same moment, the dangerous driving behavior being a driving behavior having an accident risk. Compared with determining whether the dangerous driving behavior exits according to the running status of the vehicle only, in the embodiments of the present disclosure, determining the dangerous driving behavior with reference to a more visual driving video provides more comprehensive determining bases, thereby improving accuracy of a determining result.