Remote sensing image surface feature classification method based on semantic segmentation

Abstract

The invention discloses a remote sensing image ground object classification method based on semantic segmentation, and the method comprises the following steps: S1, collecting a remote sensing image of a ground object through an unmanned plane, and constructing a remote sensing image set; S2, preprocessing the acquired remote sensing image set, constructing a sample image data set, and dividing the sample image data set into a training set and a verification set; S3, constructing a neural network semantic segmentation model; S4, inputting the training set into the neural network semantic segmentation model for training, and adjusting parameters in the training process by adopting a verification set to obtain a trained neural network semantic segmentation model; S5, inputting a to-be-identified remote sensing image into the trained neural network semantic segmentation model to obtain a ground object classification result, and realizing ground object classification of the remote sensingimage. The method solves the problems that an existing ground object classification model is large in parameter quantity, long in convergence time during training, poor in model generalization abilityand low in recognition accuracy.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a semantic segmentation-based remote sensing image ground object classification method. Background technique [0002] With the rapid development of remote sensing technology, remote sensing technology has been more and more applied to various industries and fields such as surveying, land, agriculture, transportation, forestry, water conservancy, military, etc., especially in recent years. The emergence of high-resolution remote sensing satellites, not only It greatly enriches the geographic information we can obtain, provides data support for the development of the future surveying and mapping discipline, and also provides a new direction for the development of our future remote sensing science. High-resolution remote sensing images usually contain rich detailed information of roads, waters, buildings, trees, crops, vehicles, pedestrians and other types of objects. Remo...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the above-mentioned deficiencies in the prior art, the present invention provides a remote sensing image object classification method based on semantic segmentation, which solves the problem of large parameters of the existing object classification model, long convergence time during training, poor model generalization ability and Problems with low recognition accuracy

Images