Unmanned aerial vehicle scene video target detection method based on convolutional neural network

A convolutional neural network and target detection technology, which is applied in the field of video target detection in UAV scenes based on convolutional neural networks, can solve the problems of inability to meet the real-time requirements of algorithms, slow video target detection and reasoning, and achieve power consumption. The effect of low, reduced computing load, and accelerated network inference speed

Pending Publication Date: 2022-05-31
DALIAN UNIV OF TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The present invention aims to provide a video target detection framework for unmanned aerial vehicle scenes, aiming to solve the problem of slow speed of video target detection and reasoning in the case of limited computing power of embedded platforms on unmanned aerial vehicles in the current technology, which cannot achieve the real-time performance of algorithms in actual application scenarios asked questions

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
  • Unmanned aerial vehicle scene video target detection method based on convolutional neural network
  • Unmanned aerial vehicle scene video target detection method based on convolutional neural network
  • Unmanned aerial vehicle scene video target detection method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0040] Below in conjunction with the accompanying drawings and technical solutions, the specific embodiments of the present invention will be further described.

[0043] The HeavyDet model is a MobileNet-based SSD detector. In pursuit of high accuracy, the image output

[0044] The traditional NMS is mainly used for object detection to extract a bounding box with high confidence in a picture, while suppressing the confidence

[0045] The MiniDet model is also a MobileNet-based SSD detector with a small search area as its

[0049] The target detection algorithm of this method is composed of HeavyDet and MiniDet, both based on the SSD target detection algorithm.

[0051] It can be noted that the MiniDet model may fail during its lifetime. If MiniDet misses a

[0052] Furthermore, when the camera moves rapidly, the position of the object within the scene changes drastically. For MiniDet

[0062] The labeled dataset is divided into a training dataset, a test dataset and a validation...

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 belongs to the field of video target detection in the field of computer vision, and provides an unmanned aerial vehicle scene video target detection method based on a convolutional neural network. According to the framework, information association of pedestrians between front and back frames of a video can be fully utilized to reduce a to-be-searched area and reduce the calculation load, and aiming at the problem of insufficient calculation power of an embedded platform, a tensorrt quantitative acceleration technology is used to further accelerate the network reasoning speed, so that good balance is achieved between the accuracy and the speed. According to the invention, an algorithm framework is deployed on an NVIDIA Jetson TX2 embedded platform, pedestrian target detection is carried out in an unmanned aerial vehicle scene, and the platform has the advantages of being small in size, low in power consumption, suitable for embedded application and the like.

Description

A video target detection method in UAV scene based on convolutional neural network technical field [0001] The present invention belongs to the field of video object detection (video object detection) in the field of computer vision, In particular, it relates to image classification, target detection and neural network quantification acceleration technology, specifically a convolutional neural network-based Video target detection method in UAV scene. Background technique [0002] With the rapid development of low-cost commercial UAVs, video surveillance on UAV scenes is becoming more and more important. Note. Several studies have addressed this issue from different perspectives. However, few attempts have been made on embedded platforms. This work is mainly to develop an effective and efficient drone-based drone on the Nvdia Jetson TX2 platform Human Detection Algorithm Framework. [0003] Existing detectors can be roughly divided into two-stage (e.g., faster-RCNN) ...

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): G06V20/40G06V20/17G06V10/82G06N3/063G06N3/04G06N3/08
CPCG06N3/063G06N3/08G06N3/045Y02T10/40
Inventor 卢湖川赵庆宇
Owner DALIAN UNIV OF TECH
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