Research On Leak Detection Technology Of Bag Filter Based On Digital Image Processing

Bag filter is widely used in industrial dust removal.Bag damage is the main reason affecting the normal operation of the bag filter.It is of great significance to detect the bag leakage and find the damaged bag for the efficient and stable operation of the bag filter.It also lays a foundation for automatic fill the bag leakage.This thesis proposes a detection method based on digital image processing technology to study bag leakage and broken bag positioning aiming at the difficulties of current bag filter leak detection technology,such as difficulty in maintenance and positioning of damaged bags.The bag in the bag filter studied in this paper is circular and evenly arranged.The positioning of the bag in the filter is as long as the circle bag contour on the plate is detected.In this thesis,the edge of the circle in the image is firstly extracted and two detection algorithms are proposed.The first method is to randomly sample two points in the edge detection image and detect the circle by using the gradient information of the two points and the radius constraint.The second method is to traverse each point in the edge image,and then calculate the center of the circle based on the gradient information and radius constraint of the point,and further improve the morphological operation of the circle center frequency image.Then the circular detection algorithm proposed in this thesis is compared with the detection algorithm of the Hough circle transform.The second proposed method not only detect all the circles in the image,but also the detection speed faster than the Hough circle transform.This thesis proposes to detect smoke leakage through the method of image local entropy,compares it with Background subtraction method and smoke color feature detection method.The Background sutraction method and the local information entropy detection method can accurately detect the leakage smoke.The image local information entropy method can locate the broken bag while detecting the leakage.This thesis also proposes a method for locating a broken bag using an area where leakage smoke is detected in the case of global leak detection.Through experiments,proof this method can accurately locate the position of a broken bag.