Analysis And Processing Of X-Ray Image Of Log With Defacts Based On Fractal Theory

The paper intends to analyze the typical defects which influences the log processing and to put forward a nondestructive detection to log by X-ray detecting measure with computer digital image processing and fractal Brownian random field model and fractal parameter for log X-ray image. These are to handle and analyze efficiently the detection of the hollow trunk, defect in lined skin and the log's leakage gnarl.When applying X-ray through the cross section of the log, the different absorption rates in the defect areas are shown. Generally, the passage of light in the log's defected area is higher or lower than the undefected areas. Thus by detecting the differentiation of the light, the defects of the log are seen. By placing an enlarging receptor on the other side of the log and with by feeding the image to A/D converter via digital dim light camera, a digital image is stored into the computer which is to handling the image fed with computer digital image process technology. According to the defect of different type in log, histogram balance, fast average filter, fast mediate filter, differentiation sharp, Laplace sharp operator, Sobel sharp operator, Kirsch sharp operator, etc. Regarding the circular defects such as leakage gnarl in log, this paper designed an improved Laplace's special gradient method to clarify the target image. This method is efficient in calculation, speedy treatment and can be applied extensively to detecting the log defects.Basing on the special varying characteristics of log's absorption rate and the nondestructive X-ray detection theory, the author of the paper finds the mathematical model of image between the X-ray parallel light source and the spot of light in the log. The author also defines the thick-increase and thick-decrease types of defects by finding the varying degree of log's absorption of light, thus offering a theoretical basis for detecting log's defects.The author puts forward the first time the application of edge of imageprocessing based on the Fractal theory which uses Fractal Brown's random field model to describe natural scenery. While analyzing the model, the author discovers that within small area of the image, there are statistically meaningful similarities in the gray scale, but in the image's boundary areas or margins between different images, this shape model's law is not specified as the Fractal parameter H value of the shape shows irregularities. Based on this, we can decide boundary between images. Generally speaking, a complexity in calculation of the Fractal parameter H of a image occurs as calculation work is hard, time-consuming and not satisfactory. This paper aims at, by analyzing log's X-ray image and basing on Brown's random field model, finding out a quick calculation method of Fractal parameter H. And after applying the method to calculate the log X-ray image's Fractal parameter H, joining the abnormal points, the defected boundary of the image is found. This method of searching for leakages gnarl in the log and in the X-ray images which show lined skin and hollow parts of defect areas is proved to be efficient. This is proved to be a new way in image processing technique. And at the same time, we believe it to be significant in computer pattern recognizing of defects in logs.