Study On The Separation Method Of Overlap Rods Based On Digital Image Processing

When using microscope to observe the particles, the sample would inevitably appear crossing, stacking, and so on.These are urgent problems when using image processing method to measure crossing particles. This paper study separation method of rod crossing particles.Firstly, object detected by Sobel operator edge detection algorithm which contain gaps,then the whole image of object can be obtained by dilating and filling gaps.Secondly, introduced the edge detection algorithm based on Sobel, the edge image have gaps. Corners of particle image can be detected by using a corner detection method that based on the improved curvature scale. It uses the Canny edge detector to get the binary edge map.and computes the curvature at a relatively low scale for each edge. It can pick the candidate corners set.Finally, skeleton intersection points can be formed by thinning binary image,extracting particle skeleton particle crossing points and particle corners, then the overlap particles can be separated by using crossing points.We take all kinds of image that different types of wollastonite particles.and separate the overlap rods by the separation method mentioned above, after separating overlap particles,we can calculate draw ratio of particles,and maximum and minimum value of draw ratio.