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

Hole repair model training method, device and system and hole repair model repair method, device and system

A technology for repairing models and training methods. It is applied in image data processing, instruments, computing, etc., and can solve the problems of incomplete model reconstruction, holes, and poor display effects.

Active Publication Date: 2020-12-04
REALSEE (BEIJING) TECHNOLOGY CO LTD
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] A technical problem to be solved by the embodiments of the present invention is to provide a hole repair model training and repair method, device, and system to solve the incomplete model reconstruction caused by the defects of the model reconstruction technology itself during the reconstruction process of the existing 3D model. There are holes, poor display effect and other problems

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
  • Hole repair model training method, device and system and hole repair model repair method, device and system
  • Hole repair model training method, device and system and hole repair model repair method, device and system
  • Hole repair model training method, device and system and hole repair model repair method, device and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0094] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the invention unless specifically stated otherwise.

[0095] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0096] In the embodiment of the present invention, the structural model of the room where the point is located can be reconstructed directly through an RGB color panorama by means of deep learning.

[0097] In this embodiment, by building a neural network, the neural network is trained by using massive data within the company. The neural network inputs the panoramic image and outputs the corresponding depth map. Through the panoram...

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 embodiment of the invention discloses a hole repairing model training and repairing method, device and system, and the method comprises the steps: carrying out the preprocessing of a panoramic image, and selecting a part of panoramic image to carry out the highlight or low-light processing of the panoramic image; inputting the processed panoramic image into a neural network to obtain a simulated depth map; and according to the original depth map corresponding to the panoramic image and the simulated depth map, training a hole repair model under the condition of strong light source irradiation or low-brightness irradiation. and then, performing hole repair processing on the new panoramic image by using a hole repair model. Hole repair processing also needs to refer to the self-confidence degree of the depth map, and repair is carried out by referring to the depth value with high self-confidence degree. Point cloud depth information of an existing equipment blind area is obtained through a deep learning method, and therefore the problem of model holes is solved.

Description

technical field [0001] The invention relates to the technical field of computer three-dimensional space modeling, in particular to a method, device and system for training and repairing a hole repair model. Background technique [0002] In the data collection stage of the indoor 3D model, it is generally necessary to use special equipment to collect different shooting point data (including depth data and RGB image data). After the data collection is completed, the RGBD image is obtained through the calibrated camera parameters, RGB image and depth image, so that the RGBD image can be converted into a point cloud of a single camera point in combination with the camera pose. [0003] Due to the limitation of the acquisition scheme of the device itself (the depth camera calculates the spatial information of the object surface by transmitting and receiving structured light information), in some special cases, the acquired depth is incomplete, and in some cases, the depth informa...

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): G06T5/00G06T7/50
CPCG06T7/50G06T2207/10028G06T2207/20081G06T2207/20084G06T5/77
Inventor 刘程林
Owner REALSEE (BEIJING) TECHNOLOGY CO LTD