Test Specimen Display Recognition for Mixed Display States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing test specimen display recognition devices struggle to accurately recognize test specimen displays when not all displays are lit, require template images for each type, and fail to account for changes in display states such as pointers, gauges, or characters.
Innovation Solution
A test specimen display recognition device that allows users to select and set image recognition algorithms tailored to each display type, incorporates change information, and corrects for camera angle and image distortion, enabling accurate recognition of individual displays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all test specimen displays are lit simultaneously to capture an all-lit image for recognition, then the recognition device can identify the position and type of displays, but it requires template images for each display type and cannot recognize displays that are not lit or change state
Solution Approach 1:
The patent divides the recognition process into multiple stages: first capturing an all-lit image to identify display positions and types, then capturing subsequent images to recognize individual display states. This segmentation allows the system to handle both lit and unlit displays effectively by processing recognition in discrete steps rather than requiring all displays to be lit simultaneously.
Solution Approach 2:
The system performs preliminary recognition of display positions and types using the all-lit image before capturing and processing individual display state images. This preliminary action establishes a foundation for subsequent recognition, allowing the system to identify what to look for in later images without requiring all displays to remain lit throughout the entire process.
2Ease of manufacture
If template images are prepared for each display type to enable recognition, then recognition can be performed using all-lit images, but template preparation becomes complex and recognition fails when no template exists or displays change state
Solution Approach 1:
Instead of requiring manual template preparation for each display type, the system captures actual images of the test specimen displays during operation and uses these captured images as reference templates for recognition. This copying approach automatically adapts to the specific display configurations present in the test environment, eliminating the need for pre-prepared templates and improving reliability when displays change state.
Solution Approach 2:
The system performs self-calibration by automatically capturing images of the test specimen displays and using these images to establish recognition references. This self-service mechanism eliminates the need for external template preparation and allows the system to adapt to different display configurations automatically, improving both ease of manufacture and reliability.
3Ease of operation
If image recognition is performed on displays that are not fully lit, then individual display recognition becomes possible, but position and type recognition becomes difficult without all displays lit
Solution Approach 1:
The patent segments the recognition process into two distinct phases: first using the all-lit image to identify display positions and types, then using subsequent individual display images to recognize specific display states. This segmentation allows each phase to optimize for its specific purpose without compromising the other.
Solution Approach 2:
The system performs preliminary identification of display positions and types using the all-lit image before attempting to recognize individual display states. This preliminary action ensures that even when individual displays are not lit, the system knows where to look and what to expect, maintaining position and type identification accuracy while enabling individual display recognition.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention provides a test specimen display recognition device that recognizes a test specimen display to indicate a state of a test specimen from an image of a meter panel of the test specimen captured by a camera, and the test specimen display recognition device comprises a setting information reception section that receives a selection signal indicating a type of an image recognition algorithm used to recognize the test specimen display from the image, a setting information storage section that stores the selection signal received by the setting information reception section as the setting information, and a test specimen display recognition section that acquires the setting information from the setting information storage section and recognizes an indicated value indicated by the test specimen display based on the setting information.