Optical Code Reading With Adaptive Segmentation and False-Positive Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing code reading technologies struggle to adapt to dynamic and challenging scenarios, such as varying object heights and contrasts, leading to reduced read rates due to false positives and insufficient processing time, especially in logistics applications.
Innovation Solution
A method that employs a combination of machine learning, specifically a neural network, for initial segmentation to identify code regions, followed by classical image processing to refine these regions, dynamically adjusting parameters based on the initial candidates to enhance code detection and decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image processing algorithms with manually created classifiers are used for segmentation, then small code structures can be recognized effectively, but numerous false-positive code areas are detected in challenging reading situations with many background structures
Solution Approach 1:
The system uses feedback loops where segmentation results are evaluated and used to adjust parameters for subsequent processing. The read rate and false-positive rates are continuously monitored and fed back to optimize the segmentation parameters dynamically, improving accuracy while reducing false positives.
Solution Approach 2:
The patent implements dynamic parameter adjustment where segmentation parameters are not fixed but adapt in real-time based on image characteristics and performance metrics. This allows the system to respond to varying conditions and reduce false positives while maintaining detection accuracy.
2Reliability
If a high search contrast parameter is used to ensure code areas are recognized, then low-contrast code areas may be missed, but code areas with only background texture are detected as false positives
Solution Approach 1:
The system dynamically changes the search contrast parameter based on image characteristics and performance feedback. Instead of using a fixed high or low contrast threshold, the parameter is adjusted in real-time to optimize the balance between detecting low-contrast codes and avoiding false positives from background textures.
Solution Approach 2:
The patent applies different contrast thresholds to different regions of the image based on local characteristics. Rather than using a global threshold, the system adapts the search contrast parameter locally to match the specific conditions of each code area, improving reliability while reducing false positives.
3Adaptability or versatility
If multiple code readers are installed on a single reading tunnel to capture objects from multiple sides, then objects in any orientation can be read, but device complexity increases
Solution Approach 1:
The patent makes a single code reader universal by implementing advanced image processing and parameter adaptation capabilities that allow it to handle codes in various orientations and conditions. This multi-functional approach replaces the need for multiple specialized readers, reducing device complexity while maintaining versatility.
Solution Approach 2:
Instead of adding more readers in physical space, the patent processes images in the digital domain with enhanced orientation detection and parameter adaptation. This dimensional shift from physical multiplication to computational processing achieves multi-orientation capability without increasing hardware complexity.
4Productivity
If fixed segmentation and decoding parameters are used for all objects, then optimization is only possible on average, but read rate decreases when object conditions vary dynamically
Solution Approach 1:
The system transitions from static fixed parameters to dynamic adaptive parameters that change in real-time based on object characteristics and performance feedback. This allows the system to maintain high read rates across varying conditions by continuously optimizing parameters rather than relying on average optimization.
Solution Approach 2:
The patent implements feedback mechanisms where segmentation and decoding performance is continuously monitored and used to adjust parameters for subsequent objects. This closed-loop approach enables the system to adapt to dynamic conditions and maintain high productivity across diverse scenarios.
Data Source
Figure 1~2
Figure 3~4b
Figure 5~6
AI summary
A method for reading optical codes (20) is described, comprising the steps of capturing an image, locating code regions (30) in the image, and decoding the optical codes (20) within the code regions (30). The code region locating process (30) includes an initial segmentation procedure using machine learning to identify initial candidates for code regions (30). These initial candidates are then evaluated to determine parameters for locating code regions (30) and/or decoding the optical codes (20).