Roof segmentation method, system and device based on multi-scale three-dimensional prior information

Abstract

The invention relates to the field of computer vision, in particular to a roof segmentation method, a system and a device based on multi-scale three-dimensional prior information, aiming at improvingthe accuracy and fineness of roof segmentation. The roof segmentation method of the invention comprises the following steps: firstly, multi-scale superpixel segmentation is carried out on an aerial photograph image to obtain a multi-scale segmented image layer; secondly, the pixel-by-pixel depth information based on mixed dark channel a priori, the nearest neighbor contrast information corresponding to the segmented region and the shape a priori information corresponding to the segmented region are calculated for each scale segmentation layer, and the mixed probability map of the nearest neighbor contrast information and the shape a priori information is calculated. Then, the probability map model is used to fuse the mixed probability maps of multiple scales to obtain the final probabilitymap, and the initial roof segmentation result is obtained according to the final probability map and the automatic threshold segmentation method. Finally, according to the initial roof segmentation results, the high-order conditional random field is used to generate the accurate roof segmentation results. The invention improves the accuracy and fineness of roof segmentation.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a roof segmentation method, system and equipment based on multi-scale three-dimensional prior information. Background technique [0002] Extracting buildings from aerial data has a wide range of applications in urban planning, 3D urban reconstruction and other fields. Although research in this area has made great progress in recent years, it is still very challenging to design a general and robust method. Buildings often take on different appearances due to changes in lighting conditions, reflectivity, image resolution, and image quality. [0003] Most existing building extraction methods analyze and process image features based on some settings. As man-made objects, buildings often have regular shapes, consistent color distribution, and distinct distinctions from their surroundings. Although the effectiveness of the above features has been demonstrated in many studies, the prob...

AI Technical Summary

Problems solved by technology

On the one hand, these settings are not correct for every building, on the other hand, the features extracted from the images used to support these settings often contain wrong information, and these errors are difficult to completely avoid

Images