Space target 6D attitude estimation technology based on image segmentation Mask and neural rendering

A space target and image segmentation technology, applied in 3D image processing, neural architecture, image data processing and other directions, can solve the problems of low stability and high labor and time cost, achieve low dependence, reduce labor and time costs, and improve Matching the effect of computational efficiency

Pending Publication Date: 2021-03-16
PLA PEOPLES LIBERATION ARMY OF CHINA STRATEGIC SUPPORT FORCE AEROSPACE ENG UNIV
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problems of low stability of feature extraction in existing pose estimation technology and high manpower and time cost of multi-instance high-granularity viewpoi

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
  • Space target 6D attitude estimation technology based on image segmentation Mask and neural rendering
  • Space target 6D attitude estimation technology based on image segmentation Mask and neural rendering
  • Space target 6D attitude estimation technology based on image segmentation Mask and neural rendering

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0018] specific implementation plan

[0019] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0020] Such as figure 1 As shown, the present invention provides a 6D pose estimation technology for spatial objects based on image segmentation Mask and neural rendering, including instance segmentation branch 1, neural rendering branch 2, similarity matching loss 3, and pose optimization 4.

[0021] Instance segmentation branch 1 such as figure 2 As shown, in order to achieve accurate segmentation results, the original input image is first re-cropped. In the case of ensuring the original aspect ratio, the standard image size is 1024×1024. For non-square images, zero padding is required on the short side. Such as figure 2 (a) shown. Subsequently, the anchors detection target is generated under each size feature map of the RPN, and the bounding box regression is performed. By fine-tuning the position ...

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 space target 6D attitude estimation technology based on image segmentation Mask and neural rendering in order to solve the problems that an existing method is low in featureextraction stability and high in multi-instance high-granularity viewpoint sampling labor time cost. Image segmentation Mask is taken as stable image representation and neural network differentiable rendering is taken as an attitude true value for matching calculation, attitude representation extraction and generation are carried out by introducing new image attitude representation and computer vision instance segmentation and computer graphics differentiable rendering technologies, the feature extraction stability is improved. Micro-rendering and silhouette mask binarization operations are performed on a target three-dimensional model by using a neural rendering technology, so that the rendering precision and the matching efficiency are improved.

Description

technical field [0001] The invention belongs to the field of pose estimation of a rigid object in computer vision, and in particular relates to a method for estimating the pose of a space object based on image segmentation Mask and neural rendering. Background technique [0002] Pose estimation is an important branch of many types of computer vision tasks. Its core is to calculate the position translation and orientation rotation of single or multiple types of target entities from images, and realize the expansion from two-dimensional image information to three-dimensional spatial information. Image segmentation refers to dividing each instance in the image into several non-overlapping regions through target detection, positioning, and recognition. It mainly uses low-dimensional image features and high-dimensional abstract features such as image grayscale, texture, color gradient, and edge contour. etc., clustering pixels into blocks to segment target instances, and the mask...

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
IPC IPC(8): G06K9/32G06K9/34G06K9/62G06N3/04G06N3/08G06T15/00
CPCG06N3/08G06T15/005G06V10/25G06V10/267G06N3/047G06N3/045G06F18/241G06F18/2415Y02T10/40
Inventor 杜小平杨步一方宇强郜魏柯吕潇磊张建伟曹璐柳志远倪健
Owner PLA PEOPLES LIBERATION ARMY OF CHINA STRATEGIC SUPPORT FORCE AEROSPACE ENG 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