Research On Image Acquisition System For On-Loom Fabric Defect Detection And Clear Processing Of Image

China is a big textile country.Improving the quality of textiles and the efficiency of textile equipment is an important link for China to become a powerful textile country.The defect of woven fabric is an important factor affecting its quality.At present,there are two ways to detect woven fabric defects,one is off-line detection,the other is on-line detection.Online detection can timely monitor the production process of fabrics and feedback information for process processing in real time,which can effectively improve the product grade.In online detection,the camera is compactly integrated on the loom.The loom vibration will cause the quality of the collected image to decline,which will seriously affect the subsequent image processing.This paper studies the image acquisition system of on-line cloth inspection,analyzes and improves the image quality,and lays a foundation for subsequent defect detection.The main contents and achievements of this paper are as follows:(1)Analyze the structure of loom and the weaving process of cloth,reveal the causes of cloth defects and their morphological characteristics,and determine the accuracy requirements of cloth defect detection;At the same time,the analysis of the online cloth inspection system,the design requirements of the system and the completion of the image acquisition system design;Complete the hardware selection according to the design requirements.(2)The vibration energy distribution of the loom is obtained through theoretical analysis and experimental testing,and the vibration characteristics of the loom are analyzed.The influence of vibration on camera imaging quality is analyzed from different directions.(3)Based on the image texture features,this paper proposes to use the eigenvalues of gray level co-occurrence matrix to evaluate the sharpness of cloth image.Through the experimental comparison and analysis of the effects of different methods on the sharpness evaluation of cloth image,it is proved that this method has good monotonicity,sensitivity and noise resistance.(4)Image processing method is used to clear the image blur which is difficult to be solved by hardware.The defocus blur radius is estimated by differential image autocorrelation method;By comparing the deblurring effect,anti noise and timeliness of different algorithms,Wiener filter is used to deblurring the image;Aiming at the boundary ringing of images after deblurring,a method based on boundary extension is used to eliminate the boundary ringing.(5)The image deblurring technology is applied to defect detection,the existing problems are analyzed,and the image definition threshold of the system is established.Experiments show that the method used in this paper can clear the fuzzy cloth image,and the processed image meets the definition requirements and has a good image segmentation effect.The research content of this paper can provide technical support and method guidance for the online fabric inspection image acquisition system and Image definition evaluation.