Rural building recognition algorithm based on high-resolution remote sensing image and deep learning

Abstract

The invention discloses a rural building recognition algorithm based on a high-resolution remote sensing image and deep learning. The rural building recognition algorithm comprises the following steps: 1, segmenting and calibrating a remote sensing image; 2, counting image calibration data, and segmenting the image calibration data to obtain a training set and a test set; 3, training a building recognition model: obtaining the confidence coefficient of each calibration rectangular frame by using a yov3 model, reducing the loss of the calibration rectangular frames to the minimum by using a convolutional network model rotation training mode, and obtaining a building recognition model with high test accuracy; and 4, performing identification according to the trained model. The multi-channel picture can be compressed into the rgb single-channel picture, so that resource occupation is reduced, the recognition efficiency is high, the recognition precision is high, and powerful data support can be provided for urban planning or rural construction.

Description

【Technical field】 [0001] The invention belongs to the cross field of remote sensing image analysis, artificial intelligence and architectural technology, and in particular relates to a rural building recognition algorithm based on high-resolution remote sensing images and deep learning. 【Background technique】 [0002] Land plays an important role in human production and life, and is a non-renewable resource. The population environment capacity is limited by the land area factor, and the development of human production is also deeply affected. In the process of urbanization, in many places, the understanding of urbanization is simply one-sided, and urbanization is considered to be the expansion of the city in terms of geographical area and the expansion of buildings in space. Without planning, there will be problems such as a large amount of land being occupied blindly. Improving the utilization rate of urban land has become a top priority. Traditional remote sensing recog...

AI Technical Summary

Problems solved by technology

However, this method has some disadvantages. First, multi-channel images usually take up a large space, and at the same time, the calculation is relatively more resource-intensive.

Secondly, it can only identify whether it is a house, and cannot make a specific distinction for specific elements

Images