Image detection system for fabric seam line in calendering process
Technical Field
The invention relates to an image detection system for a fabric seam line in a calendering process, which comprises image acquisition, image processing and the like of a moving fabric, belongs to the field of novel textile detection, and particularly relates to the image acquisition, processing and other technologies of embedded machine vision.
Background
In the calendering process often involved in textile enterprises, when a hot rubber roller irones a seam line of a fabric, a mark is left on the surface of the rubber roller due to sudden increase of the thickness of the seam, and the mark can be copied to the surface of the subsequently ironed fabric, so that the appearance and the flatness of the cloth surface are influenced, and even defective products are generated. Therefore, in the traditional calendering process, the lifting of the hot rubber roller needs to be controlled by manual operation to avoid the seam line, so that the quality problem is avoided. However, this method has visual errors, reaction speed errors, fatigue errors caused by continuous tension, and the like, and these subjective factors may cause excessive avoidance or insufficient avoidance of the seam line.
In recent years, with the development of image processing technology, the application range thereof is becoming wider and wider. In the textile industry, many detections related to fabrics depend on the vision of detection personnel, such as fabric defect detection, foreign fiber detection and evaluation, and the development of image processing technology, so that the problems are expected to be solved by an automatic method.
Aiming at the problems in the seam line detection in the calendering process, the invention provides an automatic detection system for the seam line of the fabric based on an image analysis technology. The digital camera is used for collecting the moving image of the fabric in the calendering process, the moving image is input into the embedded chip, whether a seam line exists in the image or not is detected, and when the seam line is detected, the rubber roller in the machine is automatically controlled to be lifted up to avoid the seam line. The implementation flow of the detection system provided by the invention is shown in fig. 1.
Disclosure of Invention
The invention aims to provide an image detection scheme of a fabric seam line in a calendering process, which is used for realizing automatic detection of the fabric seam line. The technical scheme adopted by the invention is as follows:
(1) the digital camera is used for completing the collection of images of the moving fabric, the images are input into the embedded chip through the data communication interface, the automatic detection of the seam line is completed through image analysis, when the fabric seam line is detected, the rubber roller is controlled to be automatically lifted up to avoid the fabric seam line, and an exemplary diagram of the system is shown in figure 2;
(2) inputting the collected image into a DSP development chip through a data communication line to complete embedded visual detection, wherein an AV interface is adopted as a data communication interface;
(3) judging whether a seam line exists in the image or not according to the characteristic parameters of the fabric image;
(4) when the fabric seam line is detected, the signal is transmitted to the single chip microcomputer or the PLC through the output interface so as to control the lifting of the rubber roller.
Drawings
FIG. 1 detection process of fabric seam line
FIG. 2 schematic view of a fabric seam line detection system
FIG. 3 Fabric seam line image
Detailed Description
The digital camera is arranged above the moving fabric, and it should be noted that the detection system is positioned in front of the rubber roller, when the seam line is detected, a trigger signal is output to control the lifting of the rubber roller, and the moving distance of the fabric should be less than the distance between the detection system and the rubber roller within the detection and control time.
Since the width of the seam line of the fabric is consistent with the width of the fabric, the detection does not need to shoot the image of the whole seam line, but only needs to shoot a part of the seam line. Therefore, when the digital camera is installed, the digital camera is located on one side of the moving fabric, the camera is about 20cm away from the fabric, the system provides a double-side transmission white light source, so that the seam line in the shot fabric image is clear and recognizable, and fig. 3 is the fabric seam line image shot in the actual detection system.
The image data is transmitted to the embedded chip through the data connecting line, the data line of the AV interface is adopted by the system, and the DSP (DM642) is adopted by the embedded chip.
In the DSP chip, calculating the parameters of the mean value and the standard deviation of each frame of transmitted image, and further calculating the characteristic parameters of the variation coefficient of the image, wherein the calculating method comprises the following steps:
<math>
<mrow>
<mi>CV</mi>
<mo>=</mo>
<mi>σ</mi>
<mo>/</mo>
<mover>
<mi>x</mi>
<mo>‾</mo>
</mover>
</mrow>
</math>
whereinThe mean value is the image mean value, σ is the standard deviation of the gray level image, and CV is the coefficient of variation of the image.
The CV value reflects the relative change of the image gray level, objectively reflects the gray level discrete degree of each group of anti-feathering cloth images, and can be used as a standard for detecting whether a seam line exists or not.
In the system of the present example, a CV value of 0.07 was selected as a criterion for determining the presence or absence of a seam line in the image of the fabric, and when the CV value was greater than 0.07, the presence or absence of a seam line in the image was considered. The value of the judgment basis in the specific example system has close relation with the light source selection of the system and the like, and the optimal judgment value can be obtained through example training.