High-precision registration method for multi-characteristic and multilevel infrared and hyperspectral images

A technology of hyperspectral images and infrared images, which is applied in the registration of remote sensing images and the high-precision registration of infrared and hyperspectral images. It can solve problems such as difficulties, registration software, and infrared-hyperspectral images. Good economic benefits, good versatility and practicability, and the effect of promoting wide application

Inactive Publication Date: 2015-06-03
北京市遥感信息研究所 +1
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, due to the differences in imaging mechanism and spatial resolution, the performance of the same object on infrared and hyperspectral images is very different, which brings great challenges to traditional feature matching methods.
In fact, it is very difficult even for map experts to manually and quickly calibrate some control point pairs in infrared and hyperspectral images
It is precisely because of the large difference between infrared and hyperspectral images that the current infrared and hyperspectral image registration technology is basically still in the research stage, and there is no dedicated infrared-hyperspectral image registration software.

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
  • High-precision registration method for multi-characteristic and multilevel infrared and hyperspectral images
  • High-precision registration method for multi-characteristic and multilevel infrared and hyperspectral images
  • High-precision registration method for multi-characteristic and multilevel infrared and hyperspectral images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0032] According to a specific embodiment of the present invention, corner point features are divided into active matching corner point features and passive matching corner point features, and face point features are divided into active matching face point features and passive matching face point features. Active matching angle (surface) point features are the corner (surface) point features that need to be matched on the reference image (such as infrared images), and passive matching angle (surface) point features are possible matches on the image to be registered (such as hyperspectral images) Angle (face) point feature. Active matching angle (surface) point features are a subset of passive matching angle (surface) point features.

[0033] Step S3: Match the SIFT features extracted from the low-resolution infrared image and the salient band image to obtain a matched SIFT feature pair, and use the matched SIFT feature pair and the GDBICP method to obtain a transformation model. ...

specific Embodiment approach

[0036] According to a specific implementation of the present invention, the initial transformation model is obtained by using the GDBICP (Generalized Dual Bootstrap Iterative Closest Point) method.

[0037] Step S4: On the original infrared image and the salient band image, the initial transformation model is used according to the geometric constraints provided to extract the multi-scale corner feature and surface feature based on the image block pair, and based on the initial transformation model, The scale corner point feature and the multi-scale surface point feature determine a more accurate transformation model. --Under a more accurate transformation model, the fitting error between the multi-scale angle / face point features of the infrared image and the salient band image is smaller.

[0038] According to a specific embodiment of the present invention, an iterative reweighted least square method is used to select a transformation model and obtain transformation parameters.

[0...

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 registration method for infrared images and hyperspectral images, which comprises the steps of downsampling infrared images, generating low-resolution infrared images and generating remarkable band images according to the hyperspectral images; extracting SIFT (scale invariant feature transform) features on the low-resolution infrared images and the remarkable band images and extracting angular point features and facial point features on a plurality of scales of the low-resolution infrared images and the remarkable band images; matching the SIFT features extracted from the low-resolution infrared images and the remarkable band images to obtain matched SIFT feature pairs and obtaining transformational models by the matched SIFT feature pairs and a GDBICP method; utilizing geometric constraint provided by primary transformational models to extract multiscale angular point features and facial point features based on image pairs, and confirming more accurate transformational models according to the primary transformational models, the multiscale angular point features and the multiscale facial point features; and transforming the hyperspectral images according to the transformational models to obtain transformed hyperspectral images.

Description

Technical field [0001] The invention belongs to the technical field of image processing, and specifically relates to a registration method of remote sensing images, in particular a high-precision registration method of infrared and hyperspectral images with multiple features and multiple levels. The invention can be widely applied to the registration of remote sensing images acquired by aerospace and aviation sensor platforms. Background technique [0002] Compared with visible light images, infrared images and hyperspectral images have unique advantages: infrared images reflect the temperature information of the target and can identify the working status of the target; hyperspectral images reflect the material information of the target and can identify the material of the target for use To distinguish the authenticity of the target. Therefore, the comprehensive utilization of the complementary characteristics of different sensors to fuse infrared images and hyperspectral images...

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): G06T7/00
Inventor 张秀玲霍春雷江碧涛潘春洪李京龙
Owner 北京市遥感信息研究所
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