Building Recognition Method Based on Activation Representation Replaceability in Remote Sensing Image

Abstract

The invention discloses a remote sensing image building recognition method based on the replaceability of activation expression, comprising the following steps: obtaining a remote sensing image building data set; training a common deep neural network model; calculating and identifying the independent parameters of each convolution kernel in the model Maximum Response Map; calculate the activation expression replaceability of each convolution kernel; prune the model convolution kernel according to the activation expression replaceability of each convolution kernel, and keep the small activation expression replaceability; use the pruned A deep neural network model for building recognition in remote sensing images. The method of the present invention proposes an activated expression replaceability index method for a building recognition model based on deep learning, which can quantify the replaceability of each convolution kernel on the same layer in the feature space, and the activation expression of the convolution kernel can be The lower the replaceability value, the more irreplaceable it is in the feature space, and then selectively pruning the convolution kernel, so as to effectively improve the recognition accuracy of the remote sensing image building recognition model.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image recognition, and relates to a remote sensing image building recognition method based on the replaceability of activation expression. Background technique [0002] In recent years, a large number of remote sensing satellites have been launched into the sky and have also brought a large number of remote sensing images. Remote sensing image data is increasing rapidly, including multiple remote sensing images with different spectral and spatial resolutions. These remote sensing images have brought enormous economic value. Being able to quickly extract objects such as buildings from remote sensing images can effectively help applications such as urban planning, infrastructure construction, and illegal building detection. At present, a large number of building recognition algorithms based on deep learning have emerged. However, due to the lack of understanding of the generalization of rem...

AI Technical Summary

Problems solved by technology

Many conclusions in the second option are also contradictory

However, for this method, Recht proposed that a model with good performance on a specific test set may not reflect the generalization ability of the model itself, and the correct rate based on the test set is fragile and easily changed by small changes in the data distribution.

The evaluation on the test set also has the problem of inaccuracy

Therefore, there is currently no reasonable algorithm that can correctly measure the generalization of remote sensing image recognition models.

Images