A method for radiometric calibration of color cameras using multispectral images

A multi-spectral image and color camera technology, applied in the field of radiometric calibration, can solve the problems of limited application scenarios and low precision, and achieve the effect of reducing estimation errors and loosening requirements for shooting scenarios

Active Publication Date: 2021-10-08
ZHEJIANG UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problems of the existing color camera radiometric calibration algorithm with low accuracy and limited application scenarios, and propose a method for color camera radiometric calibration using multispectral images, which can simultaneously obtain accurate camera inverse response functions and spectrally sensitive functions

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
  • A method for radiometric calibration of color cameras using multispectral images
  • A method for radiometric calibration of color cameras using multispectral images
  • A method for radiometric calibration of color cameras using multispectral images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0048] In the following, the method of the present invention is used to realize the radiation calibration of the color camera. In order to compare the calibration accuracy, first use the known camera response function and spectral sensitivity function to simulate RGB images from multi-spectral images, then use the method of the present invention to perform radiation calibration, and compare the calibration results with the known true values. from image 3 It can be seen that the two sets of data use different camera response functions, so the brightness of the corresponding RGB images is different. Wherein the GT curve represents the true value, and the Est curve represents the result estimated by the present invention. It can be seen from the comparison that the estimation results of the inverse camera response function and the spectral sensitivity function are close to the true value, and the error is small.

Embodiment 2

[0050] In the following, by means of the root mean square error, the estimation result of realizing the inverse camera response function by using the method of the present invention is measured from the perspective of quantization. from Figure 4 As can be seen in , there is a difference in the accuracy of the two estimation methods. Wherein the Single curve represents the result obtained based on the method of linearizing image boundary color distribution, and the Ours curve represents the result obtained by adopting the present invention. It can be seen that the estimation accuracy of the present invention for 201 samples is relatively high overall.

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 method for radiometric calibration of a color camera by using multispectral images, that is, solving the inverse camera response function and spectral sensitivity function of the color camera. The method includes the following steps: find the intrinsic linear relationship hidden between the pixels of the multispectral image, obtain the initial estimate of the inverse camera response function by recovering the intrinsic linear relationship from the color (RGB) image, and then use the existing data The spectral sensitivity function and the inverse camera response function coupled in the RGB image are solved alternately and iteratively. The invention uses the scene spectral information provided by the multi-spectral image, reduces the error caused by the inaccurate information used by the traditional method, improves the radiation calibration accuracy of the camera, and can obtain two calibration information of the inverse camera response function and the spectral sensitivity function at the same time .

Description

technical field [0001] The invention relates to the radiation calibration of the color camera, and can accurately obtain the inverse camera response function and spectral sensitivity function of the color camera under the condition of known scene multispectral image information. Background technique [0002] Obtaining accurate radiometric calibration information of color cameras is helpful for the implementation of many image processing algorithms, such as multispectral imaging, illumination estimation, color correction, etc. Since camera manufacturers usually do not provide this information, the camera inverse response function and spectral sensitivity function need to be estimated accurately. [0003] The camera response function converts the irradiance received by the camera sensor into the brightness of the image. In order to enhance the visual effect of the captured image, the camera response function is nonlinear in most cameras, and the role of the inverse camera res...

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): G06T7/80
CPCG06T7/80
Inventor 潘之玮沈会良
Owner ZHEJIANG 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