Since our new thinning algorithm is very quick and accurate for detecting a dynamic environment, we are trying to find more cases of the application of our digital image processing system. In this paper, we discuss our recent work on the fringe stabilizer of a long exposure holographic system; caustic fringe monitoring and computer aided analysis in SEM of HOE.

1. Set-up for the Fringe Stabilizer

This algorithm is demonstrated by processing a image of a modified Mach-Zehnder interferometer in the long exposure holographic system. A conventional fringe stabilizer, including a modified Mach-Zehnder set-up and a tiny displacement transducer mounted on a mirror of the holographic system, is applied in the long exposure holographic experiment to improve the phase shift problems of the system.

The phenomena of Mach-Zehnder interferometry is simpler to describe and can be

easily observed. Mach-Zehnder interferometry has been applied to many problems such as tiny length measurement, coherence length of a laser, thermal dynamic flow, flatness of plane plate and thickness of thin film...etc.{6} . It has proved especially useful for detecting the stability of a system.

For example, the usual procedure for checking the vibration isolated ability of a holographic table is to employ a Michelson or Mach-Zehnder interferometer on the tested table. It is apparent that a holographic system involves several phase shift problems, such as vibration; ambient temperature changes; air motion...etc. which can cause fringe motion.

The fringe stabilizer includes the photodetector to detect the fringe motion of a modified Mach-Zehnder set-up and the transducer to control the mirror position and determine the direction of mirror motion. Various equipment have been developed, based on various techniques, with two photodetectors and a PZT for detecting; computing; evaluating and compensating the phase shift by the changing the position

of the mirror. However these methods usually break down and we don't know if the photodetector or other places are damaged. It is interesting that in our system we can compare the track of the interference fringe from the original to the last.

We can also obtain the behavior of the fringe motion by the gray level of the image of the interference fringe. But here we use a 5 W old argon laser, in 457 nm wavelength with only a 0.4 W power rate, and an output power decay to 0.1 W after half an hour. The variation of the gray level does not present the motion of fringes. So thinning the fringe to find the accuracy location of the fringe is necessary. Furthermore, the thinning fringe which only depends on the gray level of the image usually induces error.

As shown in Fig.8 , the set-up of the fringe stabilizer with PZT introduced here to obtain a precise displacement data below 0.01 micro-meter digital to analog output is used to translate the displacement data and drive the PZT to push the mirror motion.

Fig.9 (A) is hard copy printed photographs corresponding to the fringes in Mach-Zehnder set-up, and Fig.9 (B) is the thinning fringe obtained by our new image processing algorithm.

At first the fringe behaved by slightly moving in the computer monitor , but when we run our computer program the interference fringes instantly froze into one position.

2. Automatically Monitor in Cracks Formation and Growth

Caustics technology is a powerful method which can be applied to the analysis of cracks formation and growth. However, the caustic curves are very complicated and vary rapidly at the instant of the material fracture. Here we develop this image processing system to analyze the fringes of the caustic{7} , and all the necessary information is obtained from the caustics' fringes by experimentally testing and computer image processing. We use a video camera to record all the caustics' image variations of the crack growth. Then we use the video equipment to reproduce the caustic image

and translate the digital data for a computer to store. We are not only interested in the region of the crack tip but also the thickness variation due to the Poisson's effect and the variation in the refractive index of the constrained material yields the fringes of the outer region in caustics. By this new thinning algorithm we obtain the caustic fringes without neglecting the other. From the outer fringes we can monitor the variation of stress and realize the stress intensity factor from the moving rate of the outer fringes.

Fig.10 (A) is the hard copy printed photographs of the caustics at the crack tip, and Fig.10 (B) is the thinning fringe obtained by our new image processing algorithm.

3. Analysis in Surface Characteristics of HOE Fringe Pattern

The surface characteristics of HOE fringe pattern are thought to have a relationship to the HOE's diffraction efficiency. But it is lubricious if we want to obtain complicated grating fringes pattern in photoresist. We need a powerful tool to measure and analyze the fringe patterns in photoresist. So we used an image processing technique to analyze the SEM photograph as the measurement of the behaviors of the HOE's diffraction efficiency and the process parameters on the photoresist plates{8-9} .

In the TV screen of SEM equipment we can see various types of surface structures appear in one HOE plate, and every part of a plate must be put into consideration for they accept almost the same process parameters. Instead of the traditional method to get these SEM pictures using a camera and film, we use an additional video camera with a multistage digital image-processing system to analyze the SEM pictures quickly, powerfully and economically. On a TV screen the image can be frozen and passed through a filter to obtain the original image process. Then the image is captured again by a video camera and soon analyzed by our new software algorithms to see sharp features in the image.

In many cases we rotate the HOE specimen to examine the shapes of the fringes from every degree of viewing angle. It's very onerous to do these work without a computer image processing aid. This work presents a simple approach to investigate the surface of a HOE having various profiles. In particular, we show that quantitative evaluation can provide a straightforward explanation of the outlines of the HOE fringes. The presentation focuses on fringes having complicated profiles which are more difficult to interpret than the profiles of the other types HOEs.

Fig.11 (A) is the hard copy printed photographs corresponding to the surface of the HOE, and Fig.11 (B) is the thinning fringe obtained by our new image processing algorithm.

nextindexback