Object Detection Switching Between Classifier and Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems for automatic machines face challenges in accurately detecting the position and posture of objects when the relative positional relationship between the object and the camera changes, leading to difficulties in tracking and controlling movable members for work operations.
Innovation Solution
An object detection device that uses a combination of a classifier trained for robust detection and a template for precise detection, switching between methods based on the camera-object positional relationship, to accurately determine the object's position and posture in real space and control the movable member accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pattern matching is used to detect objects from images, then detection speed is improved, but detection accuracy deteriorates when the camera-object positional relationship changes
Solution Approach 1:
The system dynamically switches between pattern matching and classifier-based detection methods based on the camera-object positional relationship. When the relationship is stable, pattern matching provides fast detection. When it changes, the system transitions to classifier-based detection which maintains accuracy despite positional variations, thus resolving the contradiction between speed and accuracy.
Solution Approach 2:
The system changes the detection parameter (method selection) based on the camera-object positional relationship status. By monitoring whether the positional relationship satisfies predetermined conditions, the system selects between two detection approaches, optimizing both speed and accuracy under different operational conditions.
2Adaptability or versatility
If the camera-object positional relationship changes, then adaptability is improved, but detection accuracy deteriorates
Solution Approach 1:
The system introduces an intermediary mechanism (classifier-based detection) that bridges the gap between pattern matching and variable positional relationships. When the camera-object relationship changes, the classifier acts as an intermediary that can handle diverse positions and postures, maintaining detection accuracy across different conditions.
Solution Approach 2:
The system dynamically adapts its detection method based on the camera-object positional relationship. When stability is detected, it uses pattern matching; when changes occur, it switches to classifier-based detection, thereby maintaining both adaptability to positional changes and detection accuracy.
3Reliability
If pattern matching is performed on the entire image to ensure object detection, then detection coverage is improved, but processing time increases
Solution Approach 1:
Instead of always performing pattern matching on the entire image, the system applies partial action by using classifier-based detection to identify object regions first, then applying pattern matching only within those regions when needed. This reduces processing time while maintaining detection coverage.
Solution Approach 2:
The system dynamically adjusts its detection scope based on the camera-object positional relationship. When the relationship is stable, it can use optimized search regions; when it changes, it expands coverage using classifier-based detection, thereby balancing reliability and processing time.
4Adaptability or versatility
If classifier-based detection is used to detect objects regardless of position and posture, then adaptability is improved, but detection accuracy deteriorates
Solution Approach 1:
The system dynamically selects the detection method based on camera-object positional relationship stability. When stable, it uses pattern matching for high precision; when unstable, it uses classifier-based detection for adaptability. This dynamic selection resolves the contradiction between adaptability and precision.
Solution Approach 2:
The system segments the detection process into two phases: first using classifier-based detection to identify object presence and approximate location across various positions, then using pattern matching to precisely determine position and posture when conditions allow, thereby achieving both adaptability and accuracy.
Data Source
AI summary
An object detection device detects, when a camera that generates an image representing a target object and the target object do not satisfy a predetermined positional relationship, a position of the target object on the image by inputting the image to a classifier, and detects, when the camera and the target object satisfy the predetermined positional relationship, a position of the target object on the image by comparing, with the image, a template representing a feature of an appearance of the target object when the target object is viewed from a predetermined direction.