Step 1: extract the geometric structure features of the image, that is, find the corresponding geometric structure pattern for the image to be processed.

Step 2, according to the geometric structure pattern found in step 1, select appropriate structural elements, in which the selection criteria of structural elements are firstly to show the geometric structure pattern most effectively, and secondly, the shape of structural elements is as simple as possible.

Step 3: In order to get an image that can highlight the feature information of the object better than the original image, the structural elements selected in step 2 are used to make corresponding mathematical morphological transformation on the target. If the structural elements are given appropriate variables, the geometric structure mode of the target can also be expressed quantitatively.

Step four. Through the above three steps, compared with our processing requirements, the target image will become clearer and clearer, which is more conducive to us to extract the corresponding image information.