High-resolution heterogenous remote sensing collapsed building detection method

Abstract

The invention discloses a high-resolution heterogenous remote sensing collapsed building detection method, which comprises the following steps of: firstly, constructing a unified optical-SAR ground object set based on an optical-SAR object set extraction strategy of the center of an inscribed circle of an object; then extracting high-level double-echo collapse semantic features in the SAR image based on a quantitative representation method of collapse semantic knowledge in double echoes; based on the four attributes of the area, the diagonal line, the rotational inertia and the standard deviation, a morphological attribute profile building extraction method is adopted to carry out bottom layer visual feature extraction on the optical image and the SAR image; and finally, collapsed building detection is carried out based on the improved active learning SVMs, and a collapsed building detection result is obtained. According to the method, complementary information between bottom-layer vision and high-layer semantics in multi-source data is mined, a collapsed building detection method combining post-earthquake high-resolution optics and SAR images is provided, dependence on pre-earthquake data is eliminated, and the method has important significance in timely emergency response.

Description

technical field [0001] The invention relates to a high-resolution heterogeneous source remote sensing collapsed building detection method, which belongs to the technical field of post-earthquake collapsed building detection. Background technique [0002] Timely and accurate assessment of the damage degree of buildings after an earthquake is an important part of disaster monitoring. Compared with traditional on-site survey methods, remote sensing technology adopts long-distance imaging method, which has many advantages such as timely information acquisition and is not limited by site conditions, and has become the main technical method for extracting building earthquake damage information. [0003] In recent years, the detection of earthquake-damaged buildings based on remote sensing images has mainly focused on the identification of collapsed buildings. The reason is that collapsed buildings are usually severely damaged and people are trapped, and they are the primary targe...

AI Technical Summary

Problems solved by technology

However, current deep learning methods are usually trained based on training samples from specific research areas, so the portability of the model is not yet clear; at the same time, the production and manual labeling of sample sets after earthquakes are very time-consuming and laborious. These factors have seriously restricted the application of such methods in the field of collapsed building detection.

Images