Abstract: The performance of dynamic moisture transfer for fabric, which refers to the diffusion ability of liquid water on clothing, is an important factor affecting thermal and moisture comfort of clothing. When the human body sweats, the sweat is mainly transmitted through the clothes in the form of liquid water. Therefore, it is important to explore the property of dynamic moisture transfer of fabrics to develop clothing products and to improve the clothing comfort. The traditional methods for testing the property of dynamic moisture transfer for fabrics include vertical wicking and Moisture Management Tester (MMT). With the development of computer technology, image processing technology has been widely used in the field of textile property testing such as defect detection, permeability measurement and wetting area test of fabrics for its advantages of convenience, high efficiency and accuracy. 
At present, there are still some deficiencies in the testing of dynamic moisture transfer of fabrics based on image technology: (1) the fabric wetting image taken by an CCD camera is difficult to show the capillary wetting of the edge of the wetting area, which leads to the inability to accurately obtain and identify the real edge of the wetting area of the fabric. (2) The effect of image acquisition is greatly affected by the image acquisition device, fabric color and texture, and the environment conditions. It is easy to cause image noise, which makes it difficult to extract the wetting area. In order to improve the accuracy and convenience of the measurement of dynamic moisture transfer properties of fabrics, and to extract the wetted areas on fabrics during the moisture transfer, in this paper, we proposed a method to measure the property of moisture dynamic transfer on fabrics based on the infrared image. In this method, a thermal infrared camera was used to observe the dynamic moisture transfer process of fabrics, and the temperature difference between dry and wet areas of fabrics was used to identify the wetting area. The application of infrared imaging technology overcame the interference of environmental light, fabric color and texture, and was convenient for image processing. In this paper, we also developed an automatic system to measure the dynamic moisture transfer property of fabrics based on infrared images. The system has the capacity to automatically calculate the wetting area of fabrics at any time in the dynamic moisture transfer process, and can provide accurate and reliable data for property analysis of fabrics. The program contains five modules, including the image extraction module, the image segmentation module, the image denoising module, the image binarization module and the area calculation module. The bilateral filtering algorithm was used for image denoising, and the Otsu algorithm was applied to image binarization. After importing the infrared imaging video file, the interval of the image frame was set according to sampling requirements to extract image and to segment image by the coordinates of the target wetted area. After storing the segmented image, the proposed system automatically performs image denoising and binarization, and finally converts it into the actual wetted area value of the fabric according to the ratio between pixel value and actual area and then outputs the wetted area value. 
In order to verify the accuracy of this method, six kinds of fabric samples were selected for testing, and the difference in moisture transfer property between fabrics was analyzed. The results show that the method can identify and extract the wetting area of fabrics correctly with easy operation and, high accuracy, and continuous observation of the moisture transfer process. Additionally, the results indicate that the fabric surface density and total tightness have a great impact on the initial speed of dynamic moisture transfer of the fabrics, which shows a significant negative correlation. The results are also consistent with the theoretical analysis, further proving the high reliability of the proposed method.

Key words: fabric, dynamic moisture transfer, infrared image, bilateral filtering, measurement system

摘要： 为提升织物水分动态传递性能测定的准确性，解决图像技术测定过程中织物润湿区域提取困难的问题，选用了红外热像仪对织物水分动态传递过程进行观测，构建了一款基于红外图像的织物水分动态传递性能测定系统。系统可将录制的织物水分动态传递红外成像视频按设定的参数依次进行图像提取、分割、双边滤波处理和二值化处理，并将处理得到织物水分润湿图像按照像素比例自动换算成织物实际润湿面积值输出，以自动测定织物水分动态传递性能。采用该系统测量了6种织物试样的水分动态传递性能，验证了该测定系统的准确性，并分析了6种织物的水分传递性能差异。结果表明:该测定系统能计算出织物水分动态传递过程中各个时刻的润湿面积，且结果准确性高。测试数据分析发现，织物水分传递初期的速度与织物面密度和总紧度呈明显负相关，与理论分析一致。

关键词: 织物, 水分动态传递, 红外图像, 双边滤波, 测定系统