Vehicle Camera Congestion Detection Using Grid Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies for providing information on road congestion, such as those using VICS centers, offer rough estimates and lack accuracy in identifying congestion sections and degrees, necessitating an improvement in the precision of congestion information.
Innovation Solution
A system comprising vehicles equipped with in-vehicle cameras to capture moving images of oncoming lanes, determining congestion sections and degrees based on these images, and a server that stores and provides this information to clients, enhancing the accuracy of congestion data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If congestion information is provided from a VICS center, then congestion information can be provided to users, but the accuracy of congestion section and degree identification is insufficient
Solution Approach 1:
The patent uses image recognition technology to capture and analyze actual road conditions through cameras mounted on vehicles. By copying the visual information from the road environment and processing it through AI algorithms, the system obtains precise congestion data that reflects real-time conditions, thereby improving measurement precision while preserving detailed information about congestion sections
Solution Approach 2:
The patent replaces the traditional mechanical/VICS center-based information collection system with an optical and computational system. Instead of relying on vehicle sensors reporting to a central server, the system uses image recognition and deep learning algorithms to automatically detect and analyze congestion conditions from captured images, achieving higher accuracy without losing detailed spatial information
2Measurement precision
If only head position and tail position of congestion section are acquired, then congestion information can be processed simply, but the rough congestion section cannot provide sufficient accuracy
Solution Approach 1:
The patent segments the road into multiple analysis units by dividing the image into grid sections. Each grid section is independently analyzed for congestion conditions, allowing the system to identify precise locations of congestion sections. This segmentation approach enables detailed spatial resolution while maintaining manageable processing complexity through parallel analysis of individual grid sections
Solution Approach 2:
The patent transitions from one-dimensional congestion representation (head and tail positions only) to two-dimensional spatial analysis by dividing the road visual field into grid sections. This dimensional expansion allows simultaneous identification of multiple congestion sections and their precise locations, improving measurement precision while the systematic grid approach keeps processing complexity manageable
3Measurement precision
If moving images from vehicles are used to determine congestion, then accuracy of congestion information is improved, but data collection and processing complexity increases
Solution Approach 1:
The patent implements self-service through autonomous vehicles that capture their own road condition images and automatically process them through embedded image recognition systems. Each vehicle serves itself by independently performing image capture, congestion detection, and data transmission, eliminating the need for complex centralized processing infrastructure while achieving high accuracy through distributed intelligent processing
Solution Approach 2:
The patent introduces deep learning algorithms as an intermediary between raw image data and congestion information. The AI model acts as a mediator that automatically extracts meaningful congestion patterns from complex vehicle-captured images, translating unstructured visual data into structured congestion information without requiring complex manual processing systems
Data Source
AI summary
An information processing system includes a vehicle and a server that is communicable with the vehicle. The vehicle acquires a moving image obtained by imaging an oncoming lane during traveling. At least one of a congestion section and a congestion degree of the oncoming lane is determined based on the moving image. The server stores at least one of the congestion section and the congestion degree of the oncoming lane and provides information to a client by using the stored information.