Multi-angle DR Image Inclusion Detection And Application Research Of Low Density Materials

As powder materials may introduce impurity defects such as particles or filaments during the preparation and production process,which will affect the quality and performance of the final material,non-destructive testing of powder materials is generally required.X-ray detection has the characteristics of fast detection speed and high resolution,and is a relatively common and mature non-destructive detection method.However,on the one hand,due to the X-ray imaging technology itself,the acquired Xray inclusion images usually have problems such as low contrast,weak edges,and many random noises;On the other hand,the traditional X-ray assessment of inclusion in powder materials relies on the manual method of direct measurement on the plate,which takes a long time and is inefficient,and is easily affected by the experience and fatigue of the detection personnel.At the same time,a single scan is easily affected by the spatial position of the inclusion and its own shape,which reduces the accuracy and stability of the detection.Therefore,this paper proposes a multi-angle DR(Digital Radiography)image inclusion detection method for the above problems.The main research contents of this article are as follows:(1)Aiming at the problems of large inclusion error and low stability caused by different scanning angles,this paper proposes a multi-angle DR image inclusion detection method.First,preprocess the image;then register the inclusion images at different angles to find the same inclusion;then use the image segmentation algorithm to extract the inclusions,and calculate the characteristic parameters of the inclusions based on the binary image generated by the segmentation,and select the maximum value of the inclusion size at different angles as their approximate value,at the same time,this article uses the relationship between the inclusion area and the rotation angle at different angles to predict the maximum area and rotation angle of the inclusion to improve the accuracy of detection;finally,the inclusion volume is calculated to evaluate the content of impurities in the powder material.(2)A registration method based on SIFT(Scale-invariant feature transform)algorithm is proposed for multi-angle detection of the same object.Before registration,the method of window width/window level transformation and contrast limited adaptive histogram equalization(CLAHE)is adopted in this paper to enhance the original image twice and improve the image quality.Then in the process of feature matching,this paper uses the spatial relation of matching point pairs to set the threshold value for the displacement of matching point pairs in the y direction to improve the accuracy of registration,and finally obtains a better registration effect.(3)Aiming at the weak edges of X-ray inclusions,this paper improves the deficiency of traditional image transition region descriptors,and proposes a new transition region descriptor based on the image transition region idea.The descriptor can better differentiate between weak edge and the background area,and combining the threshold segmentation to extract mixed weak edge area,and through the area fill,morphological thinning processing achieve the extraction of inclusions.The final experiment proves that the extraction method in this paper can ideally extract the inclusion area.