Movement History Extraction for Sound-Detected Surveillance Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surveillance systems often fail to detect events in areas without cameras, necessitating early detection of events and identification of relevant moving bodies.
Innovation Solution
An information processing apparatus that generates movement history information by processing images and sounds, determining event locations and times, and extracting relevant moving bodies using a related party database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If surveillance cameras are installed to detect events, then event detection capability is improved, but coverage area is limited to camera-installed locations only
Solution Approach 1:
The patent introduces sound as an intermediary medium to detect events in areas without cameras. Sound sensors capture acoustic signals from events, which are then analyzed to determine event occurrence, location, and time. This allows event detection in blind spots where cameras cannot be installed, effectively extending surveillance coverage without physical camera presence.
Solution Approach 2:
The system segments the surveillance function into two independent components: visual monitoring (camera-based) and acoustic detection (sound-based). By separating these functions, the system can deploy sound sensors in areas where cameras are not installed, creating overlapping surveillance zones and eliminating blind spots through multi-sensory coverage.
2Area of stationary object
If sound analysis is used to detect events in camera-free areas, then surveillance coverage area is improved, but event detection precision may deteriorate
Solution Approach 1:
The patent merges sound analysis results with movement history information from camera-based systems to improve event detection precision. When an event is detected by sound analysis, the system cross-references it with pre-stored movement history data of moving bodies in the area, thereby identifying specific individuals involved and enhancing the precision of event detection even in camera-free zones.
Solution Approach 2:
The system replaces direct visual observation (mechanical camera monitoring) with acoustic field analysis for event detection in certain areas. By substituting the mechanical visual surveillance system with acoustic sensing and digital signal processing, the system extends coverage to areas where physical camera installation is not feasible.
3Measurement precision
If movement history information is stored for all moving bodies, then identification accuracy of related moving bodies is improved, but data storage requirements increase
Solution Approach 1:
The system implements partial action by storing movement history information selectively rather than comprehensively. It focuses on storing data for moving bodies that are likely to be involved in events (such as those frequenting high-risk areas or exhibiting suspicious behavior patterns), while reducing or omitting storage for ordinary passing individuals, thereby balancing identification accuracy with storage efficiency.
Solution Approach 2:
The patent applies local quality by varying the storage policy based on location and individual characteristics. In high-risk areas or for specific types of moving bodies, detailed movement history is stored with high precision. In low-risk areas, storage is reduced or omitted. This localized approach optimizes storage resource allocation while maintaining identification accuracy where most needed.
Data Source
AI summary
An information processing apparatus includes a movement history generating unit that generates feature information determining a moving body being present in a target area, and position information at that time, by processing an image, also generates movement history information, of the moving body, associating the feature information and the position information with time information, and stores the generated information in a related party database, an event determining unit that detects occurrence of an event, by analyzing a sound being associated with position information and time information, and determines an occurrence location and time of the event, and an extracting unit that extracts feature information of a related moving body being at least one of the moving body related to the event, and the moving body being present at a location from which the event is witnessable, by searching the related party database by using the determined occurrence location and time.