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

Depth completion method applied to sparse map densification

A dense and sparse technology, applied in the field of robotics, can solve problems such as large amount of calculation, sparse maps, and inability to use robot navigation to achieve the effect of improving effectiveness and real-time performance

Inactive Publication Date: 2019-08-06
九天创新(广东)智能科技有限公司
View PDF1 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

On the current research issues of VSLAM, ORB-SLAM is recognized as a mature and robust VSLAM system. module, but due to its computationally intensive nature of positioning using multi-view geometry, the constructed map is quite sparse and cannot be used for subsequent robot navigation

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
  • Depth completion method applied to sparse map densification
  • Depth completion method applied to sparse map densification
  • Depth completion method applied to sparse map densification

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] In order to make the purpose, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the following The described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

[0028] see Figure 1 to Figure 5 As shown, this embodiment provides a depth completion method applied to sparse map densification, and proposes a depth completion algorithm using an adversarial neural network. Different from the existing multi-view image depth prediction algorithm, the depth completion algorithm based on the adversarial ...

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 depth completion method applied to sparse map densification. A regression model is constructed through an adversarial neural network, and the method comprises the following steps that a generator is constructed through the convolutional neural network, an L1 loss function is constructed according to a real dense depth map, a judger is constructed through the convolutionalneural network, and the L1 loss function and the loss function of the judger are weighted to obtain the loss function of the generator; the loss function of the generator is propagated back to the generator for parameter optimization, and a more accurate predicted dense depth map is obtained. The depth completion method is different from an existing multi-view image depth prediction algorithm, and the depth completion algorithm based on the adversarial neural network is beneficial for improving the effectiveness and real-time performance of depth completion, so that ORB-is realized. SLAM sparse maps are densified and applied to mobile robot visual navigation.

Description

technical field [0001] The invention belongs to the technical field of robots, and in particular relates to a depth completion method applied to densification of sparse maps. Background technique [0002] Environmental modeling using monocular or binocular vision sensors is a low-cost computer vision application field with great application prospects. Generally speaking, it is often classified as a VSLAM (Visual Simultaneous Localization and Mapping) problem in the field of robotics. . On the current research issues of VSLAM, ORB-SLAM is recognized as a mature and robust VSLAM system. module, but due to its computationally intensive nature of positioning using multi-view geometry, the constructed map is quite sparse and cannot be used for subsequent robot navigation. The densification of sparse maps usually utilizes the data association between images and the multi-view geometric relationship, so as to realize the depth prediction of each pixel of the image. [0003] Ther...

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): G06T7/50G06N3/04G06N3/08
CPCG06T7/50G06N3/084G06N3/045
Inventor 张宏黄兴鸿林旭滨陈创斌何力管贻生
Owner 九天创新(广东)智能科技有限公司