A method and system for unmanned aerial vehicle surveying and mapping based on deep learning

Abstract

A deep learning-based UAV surveying and mapping method and system, comprising the steps of: acquiring Q sensor data of the UAV, and collecting x times, wherein the sample set of the wth sensor is X; the ELO data selection module and Adboost self-enhancement The module is embedded in the neural network structure to generate the first network model; the sample set is trained to obtain the UAV surveying and mapping model, which includes up to two deep neural networks with different scales, that is, after processing by the ELO algorithm The data and / or the surveying and mapping data with different weight scales are obtained after being processed by the Adboost self-increasing module; the weights are determined through training, and the drone surveying and mapping model is obtained, and then the data collected by a single drone is surveyed and mapped. The invention embeds the ELO data selection module and the Adboost self-enhancement module into the neural network structure, realizes the matching of the preprocessed feature information and the historical surveying and mapping information, improves the accuracy of surveying and mapping and the recognition speed, and enhances the transmission efficiency, thus improving user experience.

Description

technical field [0001] The present invention relates to the technical field of computer vision, in particular to a deep learning-based UAV surveying and mapping method and system. Background technique [0002] At present, with the rapid development of UAV electronic technology, the technology related to UAV surveying and mapping is developing rapidly, and the information such as corresponding images and videos can be combined with the actual surveying and mapping scene, which greatly improves the accuracy and readability of surveying and mapping. , is a hot spot in the fields of computer surveying and unmanned aerial vehicle (UAV) in recent years. [0003] In the prior art, although there are UAV surveying and mapping methods and systems, the accuracy of surveying and mapping is low, the data transmission rate is slow, and the matching with historical data cannot be realized. There is a large demand for regular and timely updates of existing surveying and mapping data does ...

AI Technical Summary

Problems solved by technology

[0003] In the prior art, although there are UAV surveying and mapping methods and systems, the accuracy of surveying and mapping is low, the data transmission rate is slow, and the matching with historical data cannot be realized. There is a large demand for regular and timely updates of existing surveying and mapping data does not match

How to make surveying and mapping more intelligent and humanized, improve its operating efficiency and accuracy, and enhance user readability has become a new research topic, but the transmission accuracy and efficiency of existing drones are low; and it is mainly aimed at existing The processing of surveying and mapping does not involve the way of predicting current and future surveying and mapping data through historical surveying and mapping data. Therefore, an enhanced display matching technology that can increase the intelligence of surveying and mapping has become an urgent need to improve the surveying and mapping effect, thereby improving the user experience.

Images