Embedded attitude estimation method based on unmanned aerial vehicle reconnaissance image

Abstract

The invention discloses an embedded attitude estimation method based on an unmanned aerial vehicle reconnaissance image, and belongs to the field of image processing and machine vision. The method specifically comprises the following steps: acquiring an original unmanned aerial vehicle reconnaissance image data set, and performing data enhancement processing on the original unmanned aerial vehicle reconnaissance image data set; performing labeling processing on the obtained original unmanned aerial vehicle reconnaissance image data set to obtain a training data set with labels; constructing a lightweight multi-stage hourglass network, and training the lightweight multi-stage hourglass network by using the training data set; and inputting a to-be-processed unmanned aerial vehicle reconnaissance image, preprocessing the unmanned aerial vehicle reconnaissance image, inputting the preprocessed unmanned aerial vehicle reconnaissance image into the trained lightweight attitude estimation network to obtain a portrait feature map, and estimating a portrait attitude according to the portrait feature map. According to the technical scheme of the invention, algorithm performance and deployment adaptability are planned as a whole, and many problems of attitude estimation of the unmanned aerial vehicle video processing system are solved.

Description

technical field [0001] The invention relates to the fields of image processing and machine vision, and in particular to an hourglass network for estimating the embedded attitude of small ground targets for aerial video of drones. Background technique [0002] In recent years, UAVs, as an emerging combat force, have played an irreplaceable role under intelligent combat conditions. Vigorously developing UAV equipment technology is of great strategic significance for improving the combat capabilities of troops. As one of the key technologies for UAVs to perform reconnaissance and strike missions, attitude estimation technology can provide strong support for UAVs to quickly and accurately identify target intentions and travel routes. Efficient and accurate attitude estimation algorithm can effectively reduce the burden on ground operators, improve detection capabilities and rapid response combat effectiveness. [0003] The traditional UAV reconnaissance ground small target atti...

AI Technical Summary

Problems solved by technology

It mainly includes: 1) the problem that traditional attitude estimation is greatly affected by the foreground; 2) the problem that the traditional deep learning algorithm model is too large to be deployed in embedded devices; 3) the problem of slow feature extraction efficiency and poor feature fusion effect; 4 ) real-time problems in the detection process

Images