| The technology of digital image processing involves the basic theories and methods of all kinds of image information processing by using the computer. Presently, the technology of digital image processing has been widely applied to many fields, such as defense, industrial measuring, medical diagnoses, intelligent robot, transportation, communication and geology.With the rapid development of home appliances, automobile and electronic industry etc, the demand of steel strip products is growing more and more and the quality requirements are getting higher and higher. Shape quality is a key index in cold rolling strip and also an important factor in its market competition. Therefore, the technology of shape recognition plays a very important role in the process of shape rolling. At present, according to the application of the shape identifying technology in iron and steel factories, the plate shape is obtained indirectly basing on the distribution of internal force in the process of rolling steel plate. However, because of the transducer's internal factors and the external conditions, the reliability of this checking technology still can not be guaranteed completely. Then it is urgent to find a good way to check the shape of cold rolling steel effectively. Thus, it is very necessary to do some research on this technology.Basing on the current plate-profile theory and the pattern recognition theory; in term of the real situation of cold rolling sheet in a big steel factory; this paper analyzes the digital image processing technology in the application of plate-profile recognition. It also detailedly introduces the image processing technology—Image Denoising (Median filter, Mean filter, gradient prociprocal weight filter), and edge detection. In other words, they are first differential (Sobel operator, Roberts operator and Prewitt operator), second differential (Kirsch operator and Laplacian operator), LOG operator and Canny operator. All the above are studied both in theory and practice, then the results are analyzed comparatively. All the advantages and disadvantages of each operator are shown at last.Wavelet transform is a kind of time-frequency analysis method in signal processing and is becoming more and more popular in recent years. Due to its multi-scale property, localization both in time and frequency domain, sensitivity to the abrupt change of signals, and good to denoising, all these characteristics make it an ideal way to edging detection. We use static wavelet transform (SWT) upgrading to construct filter, and explore application of edge detection basing on the multi-scale wavelet transform. In this paper we make use of CCD camera, Image grabber, and Industry Control machine etc. to exploit a suit of shape imaging recognition system under Visual C++6.0 . The system gets rid of the yawp, filters the wave from the imput image, and use the Canny operator to detect the image edge while the extracting Features and design neural network which can be used as classifiers to recognize the shape edge. The device of this system is very simple and can meet precision requirement of most product line of cold rolling company, and it can also carry out untouch recognition in cold rolling successfully. The experiment result shows that five types of shape defects, which were investigated in this paper, can be recognized efficiently. |
PDF Full Text Request