Simultaneous Measurement of Spectral Distribution and Shape



1 Introduction

A present non-contact 3D shape measurement system, for example Cyberware ,which is composed of laser rangefinder and CCD color camera ,can acquire 3D shape data and the surface data (RGB color data) at a time.This system is used in the various fields . One of those is a digital archaeology .It measures historical valued object such as tea-bowl and jar ,and preserves the result as digital data. When it pays attention to color information, color information expressed by RGB data is sufficient if reconstructed 3D model is used as appreciation. However, actual color is continuous spectral distribution ,therefore if we want to do element analysis of color with reconstructed 3D model ,color information expressed by RGB data is insufficient. Particularly ,as for digital archaeology which reserves the data to posterity,it is considered that we must measure the detailed color information which can stand up to scientific investigation. Therefore, I propose a system which can measure spectral distribution and shape simultaneously with Imaging Spectrograph installed in monochrome CCD camera instead of CCD color camera. By acquiring spectral distribution corresponded with shape,we can analyze surface color with understanding the shape of the object.


2 The System Concept


This system is composed of Imaging Spectrograph installed in monochrome CCD camera ,liquid crystal projector and turntable on which a measurement object is put.


Figure1: The System Concept



2.1 Imaging Spectrograph

Imaging spectrograph is used being installed between lens and camera . The light which incidents into a slit of Imaging Spectrograph is dispersed by prism perpendicular to slit and projected to CCD camera . Consequently, we can acquire spectral information of a linear area on the object as image that has spatial axis and wavelength axis (Fig.2). In this research ,the Imaging spectrograph which can disperse the visible band 400nm 〜700nm is used.


Figure2: Imaging Spectrograph

Fig.3 Shows an acquired image with Imaging Spectrograph. Comparing with RGB three dimensional color data, color data acquired by Imaging Spectrograph has equal dimension to resolution of wavelength axis of image. Moreover, to measure circumference of an object , it is rotated by a turntable. Then the slit of Imaging Spectrograph is adjusted to the rotation axis of the turntable, and spectral distribution of the object surface is measured at each rotation angle.


Figure3: Acquired Image with Imaging Spectrograph



2.2 Measurement of Shape

Shape information is acquired with principle of light triangulation. From liquid crystal projector ,patterns are projected to object perpendicular to the rotation axis of the turntable, and camera detects reflected light from the object. As patterns ,gray code pattern is used for speedup of measurement. Each pattern divide projection space by half. Then light and darkness is detected by camera to determine direction of projection. Therefore shape is acquired from direction of projection and pixel position with principle of light triangulation.
By measuring reflected light for getting color information and reflected light for getting shape information with same measurement machinery , it would be able to correspond position between spectral data and shape data easily.


3 Result of Measurement

As measurement target, a jar shown in Fig.4 was used. In this time, measurement is done at intervals of 0.6 angles ,and in each angle 480points of 3D coordinates and spectral distribution was acquired.

Fig.5 shows result of shape measurement reconstructed as 3D model expressed by polygon mesh.

Figure4: Measurement Target , Figure5:Result of Shape Measurement



3.1 Graphic User Interface

In this research Graphic User Interface shown in Fig.6 is made . Acquired shape and spectral distribution is presented intelligibly to user using Graphic User Interface.


Figure6:Graphic User Interface

In the following , this GUI function is described.

By specifying a wavelength with usage of left scrollbar ,it can show intensity of object in that wavelength. Fig.7 shows intensity of two targets at same view point . Left one is 501nm and right one is 577nm respectively. In center of the jar , a flower design that has orange color shows low intensity in 501nm and high intensity in 577nm.


Figure7:Reflection Intensity at 501nm, 577nm

By specifying angle and height with usage of middle and right scrollbar respectively,it can show spectral distribution by graph at that position. Moreover, by pushing a right side button, a mark shown in Fig.8 appears and makes user understood where is position watched now.


Figure8:Showing a Spectral Distribution by graph



3.2 Evaluation of Shape Measurement

A cylindrical object was used as measurement object and surface was rolled by white paper shown in Fig.9 . Size of measurement object was 77mm diameter and 142mm height ,and measurement was done at intervals of 1.5 angles. Right of Fig.9 shows result expressed by points.


Figure9:Evaluation of Shape Measurement

Then central axises of both measurement object and reconstructed model were put together , and standard error of diameters was calculated. The result was 1.206mm. Consequently,if shape is grasped as object at large, taking object size into account ,error could be permissible.

4 Summary

Measurement of spectral distribution and shape was done with Imaging Spectrograph.
GUI was made ,and shape and spectral distribution was presented intelligibly to user by operating GUI.
Result of evaluation of shape measurement indicated that this system can acquire shape with permissible error.