Circular Knitting Fabric Inspection for Needle Position Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional circular knitting machines are unable to determine specific needle positions requiring repair based on fabric surface status, leading to unnecessary replacement of all needles and increased production costs due to delayed detection of defects.
Innovation Solution
A circular knitting machine equipped with a camera module and information processing unit that photographs the fabric during doffing, compares image data to identify defects, and uses an encoder to generate pulse signals for precise needle identification, allowing for immediate detection and repair of issues without requiring deep learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If quality testing is performed after knitting is finished, then fabric defects can be detected, but it is impossible to determine which specific knitting needles need replacement and all needles must be replaced
Solution Approach 1:
The patent performs quality testing during the knitting process rather than after completion. The camera module captures images of the fabric surface in real-time, allowing defects to be detected while the knitting is still in progress. This enables timely identification of problematic needles and prevents further production of defective fabric.
Solution Approach 2:
The patent establishes a feedback loop where image data is continuously captured, processed, and compared with reference images. When defects are detected, the system provides feedback about the specific needle positions, enabling targeted replacement of only the defective needles rather than all needles.
2Reliability
If all knitting needles are replaced after detecting fabric defects, then quality issues are resolved, but resource waste increases and production costs rise
Solution Approach 1:
The patent divides the knitting needle population into individual identifiable units using the encoder system. Each needle position is tracked separately, allowing the system to identify and target only the specific needles that are defective. This segmentation enables selective replacement rather than blanket replacement of all needles.
Solution Approach 2:
The patent applies quality control locally to specific needle positions rather than uniformly to all needles. By detecting defects and tracing them to specific needle positions, the system applies the replacement action only where needed, maintaining resource efficiency while ensuring fabric quality.
3Measurement precision
If feature learning is performed before production of each fabric type, then accurate defect detection is achieved, but a large amount of fabric must be knitted for learning which increases production costs
Solution Approach 1:
The patent uses a simplified comparison approach that does not require complete feature learning. Instead of performing extensive deep learning analysis, the system captures reference images during initial production and compares subsequent images against these references. This partial action approach achieves sufficient accuracy without requiring the excessive fabric production needed for comprehensive feature learning.
Solution Approach 2:
The patent performs preliminary capture of reference images during the initial phase of production for each fabric type. These reference images are then used for comparison during subsequent production runs, eliminating the need to perform feature learning from scratch each time and reducing the time and fabric required for adaptation.
4Ease of operation
If manual inspection is performed, then some defects can be detected, but defects occur before detection and defective fabric is produced
Solution Approach 1:
The patent replaces manual mechanical inspection with an automated optical inspection system using a camera module and image processing. This substitution enables continuous real-time monitoring during knitting, detecting defects as they occur rather than after production, thereby preventing further defective fabric from being produced while maintaining operational simplicity.
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
Enables instantaneous detection of defects, reducing resource waste by pinpointing problematic needles and preventing further defective fabric production, thus minimizing production costs and discarding of products.
Implementation Method 1
a camera module is provided, and the camera module photographs the fabric during doffing and generates a plurality of image data
Implementation Method 2
an encoder is provided, and the encoder generates a plurality of pulse signals when the needle cylinder rotating
Data Source
AI summary
The invention provides a circular knitting machine for prompting a knitting machine status instantaneously based on a cloth surface status of a fabric, comprising a needle cylinder, a camera module capable of photographing the fabric during doffing, an information processing unit, and an encoder. A camera lens of the camera module does not rotate with the needle cylinder, and a shooting timing of the camera lens is controlled by photographing signals. The information processing unit receives image data generated by the camera module, and compares the images, when there is a difference between the two consecutive image data on a same vertical line, a knitting machine status is prompted. The encoder generates pulse signals when the needle cylinder rotates, the encoder outputs the pulse signals to the camera module or the information processing unit, and the receiver counts the pulse signals to generate the photographing signals.