Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-case weak supervision Mars surface morphology detection method based on online learning

A technology of surface morphology and detection method, applied in the field of machine vision object detection, can solve problems such as low detection accuracy, and achieve the effect of improving detection accuracy

Active Publication Date: 2022-04-29
HARBIN INST OF TECH
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problems of low detection accuracy and convergence to local optimal solution of existing weakly supervised object detection methods, the present invention provides a multi-instance weakly supervised method based on online learning Mars Surface Morphology Detection Method

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
  • Multi-case weak supervision Mars surface morphology detection method based on online learning
  • Multi-case weak supervision Mars surface morphology detection method based on online learning
  • Multi-case weak supervision Mars surface morphology detection method based on online learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0039] Specific embodiment 1: combination Figures 1 to 5 The invention provides a multi event weakly supervised Mars surface morphology detection method based on online learning, including:,

[0040] The training set is composed of Mars perspective images; Each Mars perspective image is equipped with a terrain category label;

[0041] Setting online network includes candidate box generation unit, vgg16 network model and weak supervision and detection network;

[0042] In the candidate box generation unit, a selective search algorithm is used for each Mars prospect image to generate candidate boxes of multiple target objects or target terrain; The vgg16 network model pre trained on Imagenet is used to extract the image features of each Mars prospect image; Combining the image features and the position information of each candidate box, the full connection features of each candidate box are obtained;

[0043] The weak supervised detection network is used to detect the full connectio...

specific Embodiment

[0067] Firstly, the training samples are prepared according to the actual needs of users, and then a weakly supervised object / Terrain detector is trained according to the multi case learning (MIL) method. Then, the online optimization strategy is used to further improve the detection accuracy of weakly supervised objects and terrain, and obtain more accurate Mars surface terrain detection results. Each part is described in detail below:

[0068] First, prepare training samples.

[0069] Then the weak supervised detection network based on multi case learning is trained.

[0070] The weakly supervised detection network is an end-to-end detection method, in which in multi case learning, the label of the package is known, the label of the case is unknown, and the label of the package only indicates what kind of target exists in this image, but the location of the target is unknown. Take the Mars image as an example, each 560 × The Mars image of 500 is a package, and some patches in ...

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 multi-case weak supervision Mars surface morphology detection method based on online learning, and belongs to the field of machine vision object detection. The method aims at solving the problems that an existing weak supervision object detection method is low in detection precision and converges to a local optimal solution. Comprising the steps of generating a plurality of candidate boxes for a Mars long-range image by adopting a selective search algorithm; a VGG16 network model is adopted to carry out image feature extraction on the Mars long-range image; obtaining a full connection feature of each candidate box; in the weak supervision detection network, full connection features of candidate boxes are input, the category of each candidate box is judged through classification and detection branches, position information of the candidate boxes is scored, and finally scores of the candidate boxes are obtained by multiplying scores of the two branches and serve as case-level labels; and the K-level refined network layer takes the score of each candidate box of the multi-case learning network or the previous-level branch as supervision information, trains other optimized branches of the network, and carries out backward propagation calculation. The method is used for detecting the surface topography and the target of the Mars.

Description

technical field [0001] The invention relates to a multi event weakly supervised Mars surface morphology detection method based on online learning, belonging to the technical field of machine vision object detection. Background technology [0002] Mars is an earth like planet close to the earth in the solar system, and it is also the earth like planet most similar to the earth in the solar system; With the discovery of water, Mars is considered to be one of the most likely planets to give birth to life, and has become one of the main targets of space exploration. At present, the aerospace industry has made remarkable achievements in earth satellites and manned aerospace projects. The development of deep space exploration will be the follow-up focus, which is of great significance to scientific and technological progress and social development. Due to the complex and changeable Martian environment and often accompanied by sand and dust weather, the key technology of Mars exploratio...

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): G01V8/10G01N21/84G06N3/04G06N3/08
CPCG01V8/10G01N21/84G06N3/04G06N3/08
Inventor 张永强丁明理田瑞张印张子安王骢
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com