1996 Fiscal Year Final Research Report Summary
Realization of a Camera System with Color Constancy and Object Recognition
| Project/Area Number |
07680418
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Osaka Electro-Communication University |
|---|
Principal Investigator |
TOMINAGA Shuuji Osaka Electro-Communication Univ., Dept.of Engineering Informatics, Professor, 情報工学部, 教授 (10103342)
|
|---|
| Project Period (FY) |
1995 – 1996
|
| Keywords | color / color constancy / camera system / object recognition / spectrum / reflectance / highlight |
|---|
| Research Abstract |
1.Models for the Spectral Reflectances of Object Surfaces
Dichromatic reflection models are proposed for describing surface-spectral reflectances of a variety of materials. First we describe the standard dichromatic reflection model for inhomogeneous materials. Next the model is extended for describing the spectral reflectances of the materials for which the standard model is inadequate. The reflection characteristics are also analyzed on a chromaticity diagram. We demonstrate practical applications of the reflection models in the fields of computer graphics and color constancy.
2.Estimation of Object Color Images Using a Reflection Model
A method is described for estimating the colors of object surfaces when the measured image contains highlight and interreflection effects. We suppose that multiple objects of inhomogeneous dielectric materials such as plastics are placed close together. First, we estimate the illumination color from highlight areas in the measured image. Next, we estimat
… More
e the chromaticity of the object color of each surface from the chromaticity distribution of the image. We define a chromaticity plane. The problem of estimating object color is then reduced to finding the vertices of a polygon.
3 A Multi-Channel Vision System for Estimating Surface and Illumination Functions
We describe a set of experimental measurements and theoretical calculations designed to recover both the surface-spectral reflectance function and the illuminant spectral-power distribution from the image data. A multi-channel vision system comprising six color channels was created using a monochrome CCD camera and color filter. The spectral sensitivity of each color channel is calibrated and the dynamic range of the camera is extended to sense a wide range of intensity levels. Three algorithms and the corresponding results are introduced. First, the illuminant parameters are estimated from the sensor measurements made at multiple points within separate objects. Second the sensor responses are corrected for highlight and shading variations. The body reflectance parameters, unique to each surface, are recovered from these corrected values. Experimental results demonstrate that the illuminant spectrum and the surface-spectral reflectance functions can be recovered to within typical deviations of 1 percent and 4 percent, respectively. Less
|
