Method for rapidly visualizing multi-scale image segmentation results for large images

Abstract

The invention discloses a method for rapidly visualizing multi-scale image segmentation results for large images, which comprises the following steps: in the process of carrying out region segmentation and storage on an original image, segmenting the original image into a plurality of regions, identifying pixels in each region by using corresponding ID (identification) values so as to obtain an object ID image, and establishing an image pyramid; when a user browses, pre-loading images at the lower-resolution layer of the image pyramid; according to the operation of the user, judging the spatial resolution after the operation of the user is completed, determining the data of selected and extracted corresponding layers of the image pyramid, and according to the current view range, reading corresponding pixels in corresponding layers of the image pyramid; in the process of rendering edges of segmentation results, traversing each pixel, and judging whether the pixel is the boundary by using the most adjacent sampling; and when the user clicks some region in the current view range by using a mouse, judging the ID value of a pixel clicked by the mouse firstly, and then carrying out highlighting on the pixels with a same ID value.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image segmentation, in particular to a fast visualization method for multi-scale image segmentation results of large images. Background technique [0002] Remote sensing image segmentation technology has always been one of the key research issues in the field of remote sensing image processing. How to store and visualize the remote sensing image segmentation results is also something that every remote sensing image segmentation system must solve. At present, the commonly used solution is to store and visualize in vector form. Visualize the segmentation results of the image. [0003] Vector data is a data format that uses (x, y) coordinate pairs, coordinate strings, and closed coordinate strings to represent the position and spatial relationship of entity feature points, lines, and surfaces. It is suitable for expressing discrete spatial entity elements. In the object-oriented visualizat...

AI Technical Summary

Problems solved by technology

[0005] Due to the large amount of vector data, the complex spatial structure, and sometimes the topological relationship between objects, the speed of vector data visualization has not been satisfactory.

This is especially true for the multi-scale image segmentation results of large images. The segmentation result is a raster image with a huge number of pixels. After converting to vector data, there are a large number of point objects, which affects the speed of visualization.

Images