Research And Application Of Segmentation And Depth Extraction Technology Of Pipeline Damaged Image

The wall of metal transmission equipment is damaged owning to the accumulation of fluid erosion and corrosive for long time. The damaged metal wall causes the waste of transmitted liquid if it is not found in time. If the transmitted liquid is flammable or explosive liquid such as oil, the damaged wall may leads to weighty economic losses, meanwhile, the significant accident of casualties may be occurred when the metal wall damaged badly. Under the background, this paper researches several aspects of damaged wall image processing, including image segmentation, locating the position of damaged image, acquiring the information of damaged geometric parameters and extracting the information of depth.Image segmentation is a basic and key technology, and its aim is to divide the image into non-overlapping regions. To solve the problem of over-segmentation, this paper proposes an improved watershed image segmentation algorithm based on morphology. First of all, the paper adopts image filter to restrain the over-segmentation because of noise before image segmentation, and then uses a new evaluation of the similarity between two regions as the evaluation criteria whether regions can be merged. The algorithm prevents the phenomenon effectively that the target region merged into the background by mistake, and achieves better results being applied to the pipeline images.The technology of image depth extraction namely three-dimensional reconstruction is an important research topic in the computer vision field, plays an important and remarkable role in all areas of life, and has a broad application value. In order to obtain the depth information of inner-wall damaged region image, the paper adopts shape from shading (SFS) algorithm compared with the method that extracts depth information from defocus images, and obtains a better three-dimensional effect just needing a single gray image.