Object Detection System Using Adaptive Imaging Conditions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems face challenges in accurately counting foot passengers due to variations in camera installation positions, which restrict identification functions and require specific camera orientations for effective people counting.
Innovation Solution
An object detecting method that determines the imaging condition of an image pickup unit and selects an appropriate detection method based on that condition, using a pan-tilt-zoom control unit to adjust the camera's direction and magnification, and employing object dictionaries learned through machine learning for identifying patterns regardless of camera orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a camera is installed on the ceiling to count people passing directly below, then circular object detection can be used for identification, but the camera installation position is restricted and the system lacks versatility
Solution Approach 1:
The system implements multiple identification functions (circular object detection for top-down views, face detection for side views, and general object detection for various angles) within a single camera system. The camera can be installed in various positions (ceiling, side, or other locations) and the system automatically selects the appropriate identification method based on the imaging condition, making the system universally applicable regardless of installation position.
Solution Approach 2:
The system changes the identification parameters and methods based on the imaging condition. When the camera is installed on the ceiling, it uses circular object detection parameters; when installed on the side, it switches to face detection parameters; and for other positions, it uses general object detection parameters. This dynamic parameter adjustment allows the system to maintain high detection accuracy across different installation positions.
2Measurement precision
If a camera is installed at the side of a passage to count people crossing the field of vision, then face detection can be used for identification, but the camera installation position is restricted and the system lacks versatility
Solution Approach 1:
The system integrates multiple identification functions (face detection, circular object detection, and general object detection) into a single camera system. The camera can be installed in various positions (side, ceiling, or other locations) and the system automatically selects the appropriate identification method based on the imaging condition, making the system universally applicable regardless of installation position.
Solution Approach 2:
The system dynamically changes the identification parameters based on the imaging condition. When the camera is installed on the side, it uses face detection parameters; when installed on the ceiling, it switches to circular object detection parameters; and for other positions, it uses general object detection parameters. This parameter adaptation allows the system to maintain high detection accuracy across different installation positions.
3Measurement precision
If specific identification functions are implemented for specific camera orientations, then detection accuracy is improved for that orientation, but the system complexity increases and requires multiple identification systems
Solution Approach 1:
The system implements multiple identification functions (circular object detection, face detection, and general object detection) within a single integrated camera system. The control unit automatically selects the appropriate identification method based on the imaging condition, eliminating the need for multiple separate identification systems and reducing overall system complexity while maintaining high detection accuracy for various camera orientations.
4Measurement precision
If multiple identification systems are used for different camera orientations, then identification accuracy is improved, but the device complexity and cost increase
Solution Approach 1:
The system implements multiple identification functions (circular object detection, face detection, and general object detection) within a single integrated camera system. The control unit automatically selects the appropriate identification method based on the imaging condition, eliminating the need for multiple separate identification systems and reducing overall system complexity while maintaining high detection accuracy for various camera orientations.
Solution Approach 2:
The patent merges multiple identification functions (circular object detection, face detection, and general object detection) into a single integrated system. Instead of using separate identification systems for different camera orientations, the control unit combines these functions and automatically selects the appropriate one based on the imaging condition, reducing the number of systems needed while maintaining high identification accuracy.
Data Source
AI summary
In an object detecting method, an imaging condition of an image pickup unit is determined, a detecting method is selected based on the determined imaging condition, and at least one predetermined object is detected from an image picked up through the image pickup unit according to the selected detecting method.