Vehicle-Mounted Image Analysis for Store Status Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems face challenges in accurately detecting changes in the status of specific locations, such as out-of-business stores, due to outdated point of interest (POI) information and the difficulty in identifying specific stores among many, leading to low detection accuracy.
Innovation Solution
An information processing apparatus and method that uses vehicle-mounted devices to capture images and acquire vehicle information, allowing for the detection of changes at predetermined positions based on image analysis, including brightness values, parked vehicle counts, and pedestrian numbers, thereby improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If web information from search engines is used to detect store status changes, then coverage of geographic objects is improved, but detection accuracy of specific stores deteriorates
Solution Approach 1:
The patent segments the detection task into two levels: first, web information is used to identify candidate geographic objects (broad coverage), then image recognition technology segments and focuses on specific storefronts and signs to confirm individual store status (precise detection). This hierarchical segmentation resolves the contradiction between broad coverage and precise detection.
Solution Approach 2:
The patent introduces image recognition technology as an intermediary between web information and final detection results. Web information serves as a preliminary filter, image recognition acts as the intermediary processing layer that analyzes visual data from cameras, and the combined result achieves both broad coverage and high accuracy for specific store detection.
2Measurement precision
If manual collection of POI information is used, then data accuracy is improved, but information freshness deteriorates
Solution Approach 1:
The patent implements continuous automatic monitoring using web information crawling and image recognition technology. Instead of periodic manual updates, the system continuously collects and analyzes data from multiple sources, maintaining both high accuracy through automated verification and freshness through ongoing monitoring without manual intervention.
Solution Approach 2:
The system performs self-service by automatically collecting POI information from web sources, capturing images via mounted cameras, and processing data through image recognition algorithms. This eliminates the need for manual data collection while maintaining both accuracy and freshness through automated, continuous operation.
3Measurement precision
If image analysis from multiple parameters is performed, then detection accuracy is improved, but processing load increases
Solution Approach 1:
The patent applies partial action by selectively analyzing only the most critical image parameters (brightness value, parked vehicle count, pedestrian number) rather than processing all possible visual data. This selective approach maintains sufficient detection accuracy while significantly reducing processing load compared to comprehensive image analysis.
Data Source
AI summary
An information processing apparatus is configured to perform communication with an in-vehicle device mounted in a vehicle. The information processing apparatus includes at least one computer that is configured to acquire vehicle information including a position of the vehicle through communication with the in-vehicle device, acquire an image captured at a predetermined position with an imaging device mounted in the vehicle based on the vehicle information through communication with the in-vehicle device, and detect change in situations at the predetermined position based on the image.