Multi-Angle Image Recognition With Adaptive Camera Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image surveillance systems face issues such as blurred images due to environmental factors and inaccurate distance calculations with single-camera setups, while camera-radar systems require time-consuming calibrations and are prone to misjudgments due to obstructions.
Innovation Solution
An image recognition system utilizing multiple image capturing devices with overlapping angles and a processing device that adjusts weight values based on recognition accuracy to enhance overall recognition, reducing environmental factor influence and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single camera is used for surveillance, then the device complexity is reduced, but the recognition accuracy and detection precision deteriorate due to environmental factors and limited angle of view
Solution Approach 1:
The surveillance system is segmented into multiple independent image capturing devices, each with different angles of view. The first image capturing device captures a first image in a first angle of view, while the second image capturing device captures a second image in a second angle of view. This segmentation allows the system to overcome the limitations of a single camera while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The system transitions from a single-dimensional (single camera) view to a multi-dimensional view by incorporating images from different angles of view. The processing device integrates the first image and second image, effectively adding the dimension of angular perspective to improve recognition accuracy and reduce the impact of environmental factors.
2Measurement precision
If multiple image capturing devices with different angles of view are used, then the recognition accuracy is improved, but the device complexity and data processing requirements increase
Solution Approach 1:
The processing device merges the first image from the first image capturing device and the second image from the second image capturing device into an integrated result. The recognition module identifies the object from both images and generates a comprehensive recognition accuracy that combines information from multiple angles, thereby improving precision while managing complexity through unified processing.
Solution Approach 2:
The judgment module evaluates the comprehensive recognition accuracy and provides feedback to determine whether to adjust the weight value of the second image capturing device. This feedback mechanism allows the system to adaptively optimize its complexity by adjusting the contribution of each device based on performance, ensuring that increased complexity translates to improved recognition accuracy.
3Reliability
If camera-radar surveillance method is used, then the detection capability is enhanced, but the installation time and calibration requirements increase
Solution Approach 1:
The system employs a self-service approach where the processing device automatically integrates and processes images from multiple capturing devices without requiring external calibration during installation. The recognition module and judgment module work autonomously to evaluate recognition accuracy and adjust weight values, eliminating the time-consuming calibration process associated with camera-radar systems while maintaining enhanced detection capability.
Data Source
AI summary
An image recognition system and an image recognition method are provided. The image recognition system includes a first image capturing device for capturing a first image in a first angle of view, a second image capturing device for capturing a second image in a second angle of view, and a processing device having a recognition module and a judgment module. The recognition module identifies an object within the first angle of view to obtain a first recognition accuracy and identifies the object within the second angle of view to obtain a second recognition accuracy. The judgment module generates an overall recognition accuracy based on the first recognition accuracy and the second recognition accuracy, confirms whether the overall recognition accuracy satisfies an accuracy threshold to generate a confirmation result, and determines whether to adjust a weight value corresponding to the second image capturing device according to the confirmation result.