Unmanned aerial vehicle remote sensing river entrance discharge port identification method and device

Abstract

The invention discloses an unmanned aerial vehicle remote sensing river entrance discharge port identification method and device. The method comprises the following steps: acquiring remote sensing image features; generating a suggestion box according to the remote sensing image features and a preset identifier; after the suggestion box is accurately adjusted, classifying the river inlet outlets in the suggestion box, and obtaining classification and positioning results; and according to the classification and positioning results, generating a final identification result through a geoscience classifier. The method solves the problems that in the prior art, inherent geoscience-related information, including spatial distribution characteristics of a target andspatial correlation characteristics between the targets from a concerned target in the fields such as remote sensing and the like, except spectral information of an image cannot be utilized ; the technical problems of low efficiency and high cost due to the fact that the manual visual interpretation process needs a large amount of manpower and time to check and verify one by one images in order to avoid misrecognition and omission are solved.

Description

technical field [0001] The invention relates to the field of UAV remote sensing recognition, in particular to a method and device for UAV remote sensing river outlet recognition. Background technique [0002] With the continuous development of intelligent technology, more and more intelligent devices are used in people's life, work and study. The use of intelligent technology has improved the quality of people's life and increased the efficiency of people's study and work. [0003] At present, with the improvement of hardware (such as graphics processing unit GPU) performance and the continuous expansion of image processing technology and image recognition application fields, convolutional neural networks have been widely used in computer vision, artificial intelligence, remote sensing and many other fields involving images and graphics. in the field. The convolutional neural network extracts the feature map by convolving the image (image synthesis), and the feature map for...

AI Technical Summary

Problems solved by technology

[0006] (1) In addition to the spectral information of the image itself, it is impossible to use the inherent geoscience-related information of the target concerned in the above-mentioned fields such as remote sensing: the spatial distribution characteristics of the target, such as the distribution of the outlet of the river along the water system; The characteristics of the spatial correlation between the river outlets, such as the distribution of plants around the river outlet, the distribution of factories and other sewage units, etc.; the spatial shape of the target, such as the depression characteristics of the river outlet in the three-dimensional space relative to the surrounding

[0007] (2) After the target is extracted by the CNN network, the features of different targets are the same or similar, which affects the target recognition and classification results

[0008] (3) The spatial distribution of river outlets is relatively scattered, and the area of ​​river outlets is relatively small and hidden in large-scale remote sensing images. The manual visual interpretation process requires a lot of manpower and time to check images one by one to avoid misidentification and omissions. and verification, low efficiency and high cost

[0009] For the above problems, no effective solution has been proposed

Images