Mangrove plant automatic identification method based on unmanned aerial vehicle visible light remote sensing

Abstract

The invention relates to a mangrove plant automatic identification method based on unmanned aerial vehicle visible light remote sensing, and the method comprises the steps of carrying out the preprocessing of visible light remote sensing data collected by an unmanned aerial vehicle, and obtaining a data set needed by a training model; then generating a digital surface model of the research sampleplot through three-dimensional point cloud reconstruction; improving a SegNet semantic segmentation model, adding a group of convolutional neural networks to obtain mangrove plant physical structure features, and combining a feature training model based on visible light orthoimage extraction; and carrying out automatic classification and identification on mangrove plants by utilizing the trained model. According to the method, targeted design is not carried out according to a research sample plot, a pixel-level classification mode is adopted, universality is high, the mangrove forest physicalstructure information constraint is utilized, more comprehensive cognition of the mangrove forest is facilitated, the species identification accuracy is effectively improved, and the method can be easily applied to mangrove forest investigation and monitoring in different areas.

Description

technical field [0001] The invention belongs to the technical field of mangrove remote sensing monitoring, and in particular relates to an automatic identification method for mangrove plants based on visible light remote sensing of an unmanned aerial vehicle. Background technique [0002] Low-altitude UAV remote sensing has low cost, flexible data collection, high spatial resolution of images, and the ability to obtain image data in real time. It has certain advantages especially in small-area low-altitude fields. It is an important supplement to traditional aerial remote sensing and satellite remote sensing. In the past few years, UAVs have attracted a lot of attention in forest mapping, crop management and other vegetation monitoring, and mangrove UAV remote sensing research, as a branch of it, is in its infancy. [0003] At present, the research on mangrove plant species identification based on UAV visible light remote sensing mainly focuses on object-level species identi...

AI Technical Summary

Problems solved by technology

However, the current research in this direction mainly uses color aerial images to obtain the spectral characteristics of mangrove plants to achieve classification, and fails to make full use of information such as the physical structure of mangrove plants for a more comprehensive understanding of them.

Images