A phase moire pattern on the skin of a human body is more noisy than other interference fringe patterns. When we want to improve the resolution of the topographic contour map, we can use a sinusoidal grating ( Chang, 1991) replace the conventional rectangular grating. The sinusoidal grating which made from holographic technology is 40 lines/mm for monitoring the muscles impulsion on the pronator quadratus and 2 lines/mm for monitoring the movements of the body.

The moire fringe yields from the interference of two sets of gratings. Let the space of a grating be P1, and the space of the other grating be P2 (Fig.3(A)), and we can obtain( Kafri and Glatt 1990) :

P2 =P1 sin( -)/sin

From the above equation , we can solve the inclination slop rate of the moire fringes.

tan =P1 sin /(P1cos-P2)

The value of the inclination slop rate of the moire fringe tan depends on the relative value of P1 and P2, This provides a method to measure the angle between two gratings.

When is small, we obtain sin =, cos =1 , and then simplify the above equation:

tan = P1/(P1-P2)

The simplest moire setup comprises a light source, that projects a shadow of a phase grating at an angle alpha to the beam direction onto a 3 dimensional object and a distorted grating is observed. A moire is formed by superimposing the undistorted grating with a projecting of the distorted one to obtain an equal height fringe moire pattern. A CCD camera is placed perpendicular to the grating and recorded to analyze this moire information. From the video camera and image card we obtain two sets of fringe with different spatial frequency. The computer software resolves the low spatial frequency grating and overlooks the high spatial frequency one. Here we can use a simple algorithm to separate two types of fringes and from the variations of the gray level we can determine the phase within each fringe.

Let h(x,y) represent the height variation function of the tested object, then the amount of distortion is proportional to h(x,y)/ tan (Kafri and Glatt 1990). In other words, this represents between successive fringes a topographic contour map of the object with a height increment is (Fig.3(B)):

The digital image system consists of a CCD camera to scan the fringe pattern, and is connected to a frame grabber. The frame grabber is a video digitizer which can digitize video signals from the CCD camera in real time and store the digitized image into the on-board frame memory. The image also can be displayed simultaneously on the analog video monitor. The frame grabber is a plug-in card for the IBM PC/XT or AT. It has 256K bytes of frame memory organized as 512times 512times 8 bits. The frame grabber digitizes the incoming video signal to 8 bits resolution at a rate of 30 frames per second. With an 8-bits A/D converter, the input video signal is quantitized to 256 possible intensities or gray levels. So each pixel in the frame memory may take a value between 0 and 255, with the value 0 corresponding to the black level and 255 to the white level.

Many papers present the entire analysis of moire fringe patterns by using a digital image processing technique (Chang et al., 1994; Agin, 1976) to calculate the fringe order and produce multiform image output. A new methodology, developed for data thinning fringes of moire patterns, is utilized to extract the fringe skeletons from the original pattern. The method is applicable to any orientation of the fringe. The various steps of the algorithm are explained by processing a test image in the PC-based image processing system. According to this method, we can analyze the fringes of a moire image, and successfully extract the skeleton of the fringe, and the thinned pattern preserves the connectedness and the shape of the original pattern.

We process the moire image with the following steps:

1. Normalizing and low pass filtering.

2. Image inversion and subtraction.

3. Thinning the moire fringe along two directions.

4. Compensate the lost point( a point which gray level is the same as background but all its neighborhood pixels are object) and cut the short line( a line which include less than 5 pixels) (Chang et al., 1995).

5. Distinguish the fringe order; determine the phase within each fringe and calculate the height of the dynamic profile of the target.

6. Subtract the height of the original profile of the target and construct the three dimensional figure of the variation of the profile of the target( Lin and Chang, 1995).

index