Urban intelligent extraction method, system and device based on remote sensing image scene segmentation

Abstract

The invention discloses an urban intelligent extraction method, system and device based on remote sensing image scene segmentation. The method comprises the following steps: 1, segmenting a high-resolution remote sensing image according to road network data, and generating a block-level scene image for deep learning scene classification; step 2, initially dividing and marking urban and non-urban samples according to government locations and image features of all regions of a city, and generating a sample set containing a training set, a verification set and a test set; 3, selecting artificialfeatures, fusing the artificial features with the convolutional neural network and the features thereof, and classifying urban block scenes; 4, constructing a pyramid grid according to the scene classification result, and judging the urban internal block plaques, the surrounding plaques and the grid classification result; and step 5, iterative training and scene classification are performed on thesample set until all sample categories are determined and an urban area range is obtained, so that the effect and precision of urban area extraction are ensured.

Description

technical field [0001] The present invention relates to the technical field of remote sensing image processing, in particular to a method, system and equipment for intelligent urban area extraction based on high-resolution remote sensing image scene segmentation. Background technique [0002] With the rapid advancement of urbanization, while promoting economic development, the urban spatial pattern has become increasingly complex, facing severe situations such as tightening resource constraints, serious environmental pollution, and ecosystem degradation. In order to achieve sustainable development, cities must first grasp the status quo of urban development, rationally adjust urban layout, and control the phenomenon of blind urban expansion. Urban expansion monitoring is an important basis for scientifically guiding urban layout and realizing sustainable urban development. Urban area extraction is an important technical means for urban expansion monitoring. Compared with lo...

AI Technical Summary

Problems solved by technology

However, it is very difficult to design a good artificial feature, and usually has specific application scenarios, poor applicability, and difficult to deal with the complex situation of the city

Using deep learning to extract urban areas is accurate and effective, but it requires a large number of labeled samples for training to achieve good results, and building a large and comprehensive sample set requires a lot of manpower and material resources

Images