3D target detection method of monocular view based on convolutional neural network

A convolutional neural network and target detection technology, applied in the field of 3D target detection in monocular view, can solve the problem of missing depth information and achieve high resolution, high accuracy, and high accuracy

Active Publication Date: 2020-07-03
ZHEJIANG UNIV
View PDF7 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the perspective projection of the monocular camera, the depth information is missing, and the positioning of the 3D target is a big challenge.

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
  • 3D target detection method of monocular view based on convolutional neural network
  • 3D target detection method of monocular view based on convolutional neural network
  • 3D target detection method of monocular view based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0049] The method of the present invention will be further described below in conjunction with the accompanying drawings.

[0050] Such as figure 2 As shown, a 3D object detection method based on convolutional neural network monocular view, its specific implementation steps are as follows:

[0051] Step (1). The input image is a monocular view collected by a vehicle camera;

[0052] Step (2). The training sample is divided into a training set and a test set, and the training set sample is put into a convolutional neural network and trained by backpropagation. The test set samples are used to test the generalization ability of the model.

[0053] Step (3). Centralize and standardize the R, G, and B channels of the input image, that is, subtract the mean value obtained from the statistics on the training set, and then divide by the standard deviation:

[0054] X'=X-X mean

[0055] x s =X' / X std

[0056] Among them, X is the image to be preprocessed, X mean is the mean ...

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 discloses a 3D target detection method of a monocular view based on a convolutional neural network. 3D targets mainly detected in an automatic driving scene include but are not limited to automobiles, pedestrians, bicycles and the like. The method mainly comprises the following steps: firstly, training network parameters by using a training set prepared in advance; secondly, in the prediction stage, a monocular image collected by a vehicle-mounted camera is preprocessed and then input into the trained convolutional neural network, and the 2D frame, the actual size and the depth of a target are predicted; and finally, solving the position of the 3D target in the 3D space through the projective geometry camera model. The 3D target detection method based on the monocular view provided by the invention has relatively high precision in a test data set, and has good accuracy and robustness.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a 3D object detection method based on a convolutional neural network based on a monocular view. Background technique [0002] Vision is the main source of information that people rely on when they perceive the surrounding environment while driving a vehicle. Human vision has evolved over a long period of time, and has a good perception of the surrounding environment, and can easily identify the surrounding target objects and perform positioning perception on the surrounding target objects. And computer vision technology just wants to endow computers with the functions of human visual recognition and positioning. Through complex image calculations, the computer is able to identify and locate the target object. [0003] In recent years, autonomous driving has received great attention both in industry and academia. The purpose of autonomous driving is to replace human dr...

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): G06T7/70G06K9/00
CPCG06T7/70G06T2207/20081G06T2207/20084G06T2207/30252G06V20/56G06V2201/07
Inventor 丁勇罗述杰李佳乐孙阳阳周一博
Owner ZHEJIANG UNIV
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