Overlapping Parking Cameras for Redundant Raised-Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current parking lot monitoring systems lack efficient detection of raised objects within driving paths, leading to potential collisions and inefficient operation due to limitations in sensor redundancy and error compensation.
Innovation Solution
A method using at least two spatially distributed video cameras with overlapping fields of view, where analysis is performed both internally within the cameras and externally by a separate computing unit, allowing for dual-function video camera operation and redundant calculations to detect raised objects effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If analysis is performed only externally by a separate computing unit, then processing power is sufficient, but system complexity increases and response time increases due to data transmission
Solution Approach 1:
The analysis function is segmented between the video camera's internal processor and the external computing unit. The video camera performs initial analysis locally to detect raised objects, while the external computing unit performs additional verification analysis. This segmentation allows the system to distribute processing tasks, reducing the complexity of any single component while maintaining sufficient overall processing power.
Solution Approach 2:
The video camera performs preliminary analysis of video images internally before transmitting data externally. By conducting initial detection and filtering of raised objects at the source, the system reduces the amount of data that needs to be transmitted and processed externally, thereby decreasing system complexity and improving response time while retaining adequate processing capability.
2Reliability
If only one video camera is used, then device complexity is low, but reliability is insufficient due to lack of error compensation
Solution Approach 1:
Multiple video cameras provide feedback mechanisms for error compensation. When one camera detects a raised object, the system cross-verifies with other cameras in the network. This feedback loop allows the system to compensate for errors or failures in individual cameras, improving overall reliability without requiring a proportional increase in the number of cameras.
Solution Approach 2:
The patent merges the functionality of multiple video cameras into a unified monitoring system where cameras are spatially distributed but operate as an integrated network. By combining their fields of view and analysis capabilities, the system achieves enhanced reliability through redundancy while minimizing the total number of cameras needed compared to fully redundant standalone systems.
3Reliability
If analysis is performed both internally and externally, then detection reliability increases through redundancy, but processing time increases
Solution Approach 1:
The internal analysis in the video camera performs preliminary detection of raised objects before external verification. By pre-processing and identifying potential raised objects at the source, the system reduces the scope of external analysis needed, thereby maintaining high detection reliability through redundant verification while minimizing additional processing time.
Solution Approach 2:
The system implements a skip mechanism where if internal analysis confidently identifies a raised object, the external verification process can be expedited or selectively applied. This allows the system to maintain high detection reliability through redundant analysis while reducing processing time for clear cases, avoiding unnecessary full verification cycles.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The invention relates to a method for detecting a raised object located within a parking area by means of at least two video cameras which are placed at a distance from each other in the parking area and the visual ranges of which overlap in an overlapping sector, said method involving the following steps: having the video cameras capture video images of the overlapping sector; analyzing the captured video images in order to detect a raised object in the captured images, the analysis being performed both video camera-internally by at least one of the video cameras and video camera-externally by at least one computing unit that is different from the video cameras. The invention further relates to a corresponding system, a parking area and a computer program.