Remote sensing image earthquake damage building identification method based on decision tree and feature optimization

Abstract

The invention provides a remote sensing image earthquake damage building identification method based on a decision tree and feature optimization, and aims to solve the problems and limitations existing in feature modeling and random forest classification of earthquake damage buildings only depending on post-earthquake remote sensing images under the condition of lack of pre-earthquake reference information. Firstly, a potential building object set is extracted in combination with image segmentation and non-building screening rules; on the basis, an adaptive decision tree quantity extraction strategy based on classification accuracy curve fluctuation discrimination is provided; meanwhile, under the guidance of the characteristic importance index, three types of characteristics of spectrum,texture and geometrical morphology are screened to obtain a representative earthquake damage characteristic set; and finally, the earthquake damage building is identified based on the constructed optimized random forest model. Experiments on four groups of different remote sensing images show that the method shows excellent performance in earthquake damage building identification in a complex scene after an earthquake, and the total precision can reach more than 85%.

Description

technical field [0001] The invention discloses a remote sensing image earthquake-damaged building recognition method based on decision tree and feature optimization, which belongs to the technical field of image recognition. Background technique [0002] As a serious natural disaster, earthquakes are often accompanied by huge casualties and economic property losses. Timely and accurate identification of earthquake-damaged buildings after the earthquake is of great significance for rapid assessment of the disaster situation, emergency rescue response, and post-disaster reconstruction. Compared with the traditional post-earthquake manual field inspection method, the identification of earthquake-damaged buildings based on remote sensing images has the advantages of rapid data acquisition and wide coverage, and has become an important technical means in post-earthquake emergency response. [0003] With the continuous development of satellite and sensor technology, the wide appl...

AI Technical Summary

Problems solved by technology

Nevertheless, there is no clear standard for the number of decision trees in RF theory, and the usual manual assignment method is not only easily affected by subjective factors and falls into local optimum, but also reduces the degree of automation of the classification process.

Images