Wire Marking Identification Using Angled Lighting and CNN Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies lack effective solutions for automatically identifying alphanumerical wire markings in industrial settings due to the small size of wires, curvature, bending, and printing defects, which lead to difficulties in recognizing and reading the markings, resulting in time-consuming and error-prone manual processes.
Innovation Solution
A device comprising a camera, microcomputer, lighting system, and deep convolutional neural networks is designed to identify wire markings by processing images, recognizing gaps and characters, and using a maximum likelihood method to correct misidentifications, ensuring optimal illumination and positioning to minimize reflections and shadows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated image recognition systems are used to identify wire markings, then productivity and consistency are improved, but the system fails to accurately recognize markings due to wire bending, twirling, print tilting, blurring, and rubbing off
Solution Approach 1:
The system performs preliminary actions by capturing multiple images of the wire from different angles and positions before attempting recognition. This allows the system to have multiple candidate images to work with, increasing the probability of finding a clear, recognizable marking despite wire deformation or print defects.
Solution Approach 2:
The system implements feedback mechanisms where recognition results are continuously evaluated. When a marking is not clearly recognized, the system uses feedback signals to trigger additional image captures or adjust processing parameters, creating a closed-loop system that improves reliability through iterative refinement.
2Reliability
If multiple images are captured to ensure accurate recognition, then reliability is improved, but the time required for identification increases
Solution Approach 1:
The system applies partial action by capturing a limited number of images strategically positioned to cover the most likely locations of clear markings. Rather than exhaustively capturing all possible views, the system captures enough images to achieve reliable recognition in most cases, balancing time consumption with accuracy.
Solution Approach 2:
The system changes parameters such as camera angle, lighting conditions, and image processing thresholds dynamically based on the specific wire and marking being identified. This allows the system to optimize the number and quality of images needed for each identification task, reducing overall time while maintaining reliability.
3Measurement precision
If deep convolutional neural networks are used for image analysis, then recognition accuracy is improved, but device complexity increases
Solution Approach 1:
The system introduces intermediary components such as pre-processing modules that prepare images before they reach the neural network, and post-processing modules that refine recognition results. These intermediaries simplify the core neural network's task while maintaining high accuracy, effectively distributing complexity across multiple specialized components.
Solution Approach 2:
The system segments the complex recognition task into distinct stages: image capture, pre-processing, neural network analysis, and result validation. Each stage handles a specific aspect of the problem, allowing the use of specialized algorithms for each step while keeping the overall system manageable and maintainable.
4Ease of operation
If manual identification of wire markings is performed, then flexibility is maintained, but productivity decreases and error rates increase
Solution Approach 1:
The automated system is designed to be self-sufficient in handling the majority of identification tasks without requiring manual intervention. It automatically captures images, processes them through neural networks, and generates recognition results, freeing operators to focus on higher-value tasks while maintaining high productivity and consistent accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution enables automated and accurate identification of wire markings, reducing assembly errors, lowering manufacturing costs, and improving quality assurance by ensuring consistent and efficient recognition of wire markings in industrial environments.
Implementation Method 1
above the lead at an angle of not less than and not more than 20 degrees to the axis of the lead of a wire there is a set of lighting consisting of at least two pairs of LED headlights
Implementation Method 2
a camera, a microcomputer, a lighting system and a monitor according to the invention is characterized
Data Source
Figure 1
Figure 2
Figure 3a~3b
AI summary
The subject of the invention is the method for identifying wire markings in which digital images of a wire are taken, images are processed, image characteristics are separated and analysed using a device having the housing (1), the upper wall (3) of which is equipped with a monitor (4) 5 inclined at an angle of 12 to 18 degrees with respect to the horizontal plane of the device, inside the housing (1), beyond a camera (12) observation field, connected to the camera (12), there is a microcomputer (13) with a power supply (14), a signalling device (15) and the system (16) to control peripheral devices such as: driving lighting set, monitor (4), signalling device (15) and 10 sensors (11), where to the bottom part of the side walls (2a, 2b, 2c, 2d) of the housing (1) a longitudinal panel (5) is attached, above the panel (5) parallel to it and symmetrically in the camera (12) observation field, and perpendicular to the camera observation axis (01) there is a wire lead (6) containing the right and left lead channel (7a, 7b) and the right and left lead bed (8a, 8b), on 15 the edge of each lead bed ( 8a, 8b) at least one sensor (11) is placed, and above the lead (6) at an angle of not less than 10 and not more than 20 degrees to the axis of the lead (02) of a wire there is a set of lighting consisting of at least two pairs of LED headlights (10) located above each lead channel (7a, 7b), and at its end, symmetrically with respect to each lead bed 20 (8a, b) and symmetrically with respect to the camera observation axis (01), while the distance of the camera (12) over the axis of the lead (02) of a wire is selected so that the observation field of the camera (12) covers the full length of the wire lead (6).