Multi-scale remote sensing image target recognition system and method based on deep learning

Abstract

The invention provides a multi-scale remote sensing image target recognition system and method based on deep learning. The system adopts a serviced technical architecture, performs containerized packaging on services, and comprises a sample data acquisition service for acquiring sample marking data in a remote sensing image and acquiring image tiles corresponding to the sample marking data to obtain a sample data set; a target recognition model training service used for inputting part of sample data in the sample data set into a deep learning platform for model training to generate an optimalprediction model; and a target identification and prediction service used for inputting a remote sensing image to be identified into the optimal prediction model for target identification to obtain atarget identification result. In this way, model training and target recognition tasks can be supported on a cloud platform, system upgrading and migration deployment are facilitated, the engineeringapplication level of remote sensing image target recognition is improved, and engineering application of remote sensing image ground object automatic interpretation is easy to achieve.

Description

technical field [0001] Embodiments of the present invention generally relate to the technical field of remote sensing, and more specifically, relate to a deep learning-based multi-scale remote sensing image target recognition system and method. Background technique [0002] With the rapid development of spatial information technology, the spatial resolution, time resolution and spectral resolution of remote sensing images have been greatly improved; how to use remote sensing images to extract target information has always been a thorny and important issue in the field of image vision. In recent years, with the theoretical breakthrough of deep learning technology, remarkable achievements have been made in the field of image vision. The characteristics of remote sensing image target recognition using deep learning technology are as follows: First, the separation of model training and model application tasks has been realized. Model training The trainer performs various target ...

AI Technical Summary

Problems solved by technology

[0003] In the engineering application of remote sensing image target automatic recognition, the use of deep learning technology has the following problems: 1. There are many deep learning training platforms, and the end-to-end platform for remote sensing image target recognition training tasks is missing; 2. Deep learning model training is computationally intensive. Type tasks require a single machine with a high-performance independent GPU, or even use a GPU c

Images