Deep inverse reinforcement learning-based target detection method in unmanned aerial vehicle aerial video based on deep inverse reinforcement learning

A technology of reinforcement learning and target detection, which is used in computer parts, instruments, biological neural network models, etc.

Active Publication Date: 2019-10-11
INST OF ELECTRONICS ENG CHINA ACAD OF ENG PHYSICS +1
View PDF8 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The object of the present invention is to provide a target detection method in the unmanned aerial vehicle video of the deep inverse reinforcement learning that can solve complex tasks and reward reward delay

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Deep inverse reinforcement learning-based target detection method in unmanned aerial vehicle aerial video based on deep inverse reinforcement learning
  • Deep inverse reinforcement learning-based target detection method in unmanned aerial vehicle aerial video based on deep inverse reinforcement learning
  • Deep inverse reinforcement learning-based target detection method in unmanned aerial vehicle aerial video based on deep inverse reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0083] A target detection method in the unmanned aerial vehicle aerial video of a kind of depth inverse reinforcement learning, it is characterized in that: comprise the following steps at least:

[0084] Step 1. Establish a deep inverse reinforcement learning model;

[0085] Step 2. Model strategy iteration and algorithm implementation;

[0086] Step 3. Selection and optimization of key parameters of the model.

[0087] Described step 1 comprises following specific steps:

[0088] Step 1, establish a deep inverse reinforcement learning model

[0089] Randomly select the video clips that have acquired the target during the training process as the model input. Under the premise of n-frame association detection, define the minimum number of frames of the video clip as n+1, and the maximum number of frames of the video clip does not exceed 2n, such as the formula (1 );

[0090] n+1≤FramCount≤2n (1)

[0091] Such as figure 1 As shown, the video clips in the obtained target a...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a moving target detection technology, in particular to a deep inverse reinforcement learning-based target detection method in an unmanned aerial vehicle aerial video, which ischaracterized by at least comprising the following steps: 1, establishing a deep inverse reinforcement learning model; 2, performing model strategy iteration and algorithm implementation; 3, selecting and optimizing key parameters of the model; and 4, outputting a moving small target detection result. According to the target tracking method in the unmanned aerial vehicle aerial video, complex tasks can be solved, and award return delay can be avoided.

Description

technical field [0001] The invention relates to a moving target detection technology, especially a target detection method in a UAV aerial video of deep inverse reinforcement learning. Background technique [0002] In the field of UAV aerial video target detection, the airborne camera is usually at a high-distance shooting angle during the mission, and the target of interest in the aerial video usually has the characteristics of small targets, degraded target features, and slow motion. It is difficult to detect and extract moving objects. However, there is a strong correlation between the corresponding moving target sets between adjacent sequences of aerial video, including target center position, size information, target pixel features, etc. The time difference algorithm proposed by Sutton is used for target detection. There is a slight deviation in the process of global motion compensation for the background. Some abnormal points such as bright spots or obvious edges appe...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/32G06K9/62G06N3/04
CPCG06V20/13G06V10/25G06N3/045G06F18/214Y02T10/40
Inventor 刘友江周劼秋勇涛孙伟闫达帅杜川
Owner INST OF ELECTRONICS ENG CHINA ACAD OF ENG PHYSICS
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap