Unlock instant, AI-driven research and patent intelligence for your innovation.

A 3D object detection method based on monocular image

A target detection and 3D technology, which is applied in image analysis, image enhancement, image data processing, etc., can solve the problems of adding noise and affecting the performance of the detection framework, so as to improve the detection accuracy and ease the difficulty of training

Active Publication Date: 2022-06-07
WUHAN UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This self-network-dependent method will potentially add noise, which will affect the performance of the entire detection framework.

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
  • A 3D object detection method based on monocular image
  • A 3D object detection method based on monocular image
  • A 3D object detection method based on monocular image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0048] The technical solutions of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0049] like figure 1 As shown, the present invention provides a 3D target detection method based on a monocular image, comprising the following steps:

[0050] Step 1, generate 2D and 3D anchor frame templates. By clustering the 2D real boxes of the labeled dataset in advance, K anchor boxes of different sizes and scales are obtained, which are defined as [x p ,y p ,w 2D ,h 2D ], x p ,y p Represents the center coordinate point of the 2D anchor box, w 2D ,h 2D Represent the width and height of the 2D anchor box, respectively. The 3D anchor box is defined as [x p_3D ,y p_3D ,z 3D ,w 3D ,h 3D ,l 3D ,θ 3D ], w 3D ,h 3D ,l 3D Represent the size of the 3D box, namely width, height, length; θ 3D Indicates the rotation angle yaw of the 3D frame along the y-axis, that is, the yaw angle. Its specif...

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 3D target detection method based on a monocular image. First, make full use of the labeled data set and the calibrated camera projection matrix to obtain 2D and 3D prior anchor frames; use the pre-trained deep convolutional network as the basic network to extract the semantic information in the image; use three-way parallel volume The multi-task loss function is used to balance the losses of different tasks and different components in the same task to obtain accurate 2D and 3D frames. The invention realizes 3D target detection only through monocular images, has low cost and good effect, and can be widely used in the field of unmanned driving.

Description

technical field [0001] The invention relates to a technology in the field of unmanned target detection, in particular to a 3D target detection technology based on monocular images. Background technique [0002] At present, the development of image-based 2D target detection technology is relatively mature. Compared with 2D target detection tasks, 3D target detection tasks are more difficult, more complex, and of great significance. 3D target detection plays a very important role in the field of unmanned driving. It can not only know the size, position, and direction of movement of the target, but also understand the current state of the target, and at the same time, it has the ability to predict, track, avoid obstacles, and plan paths. Supporting role. Therefore, an accurate 3D target detection algorithm is urgently needed in the field of unmanned driving to ensure the safety and reliability of unmanned driving. [0003] The existing 3D target detection technology mainly re...

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 Patents(China)
IPC IPC(8): G06V20/59G06V10/80G06V10/82G06T7/55G06N3/08
CPCG06T7/55G06T2207/20081G06T2207/20084G06V20/59G06V2201/07G06F18/253
Inventor 李明樊一帆曹晶
Owner WUHAN UNIV