Computer Holographic Measurement of Machine Parts by Laser Interferometry

(1)为了测量工件的物面偏差，研制了一种新型的干涉仪，该干涉仪采用CCD图像传感器代替原干涉仪中的全息图。CCD图像传感器的引入提供了稳定的测量。(2)每个干涉条纹图到参考和物镜表面的亮度数据被转换为不连续的相位差。本文还提出了将计算机全息技术应用于数据处理的设想。该技术使得相位能够有效地级联。(3)CCD图像传感器上的干涉条纹图案包括厚密度的条纹，并且它还包含相当大的噪声量。即使用上述技术解决了相邻条纹之间的图案连接等问题，也仍然存在。针对这一问题，在实验中介绍了在参考光路中插入凹透镜、采用靶面形状和采用两个不同波长光源的外差法等减小干涉条纹的方法。这些方法成功地处理了从亮度数据到形状偏差的数据。(4)用相移法从干涉条纹图的亮度数据计算相位差时，需要取多少幅干涉条纹图并不明确。从使用一些平面试件的一系列实验中可以看出，6-8张图案就足够了。这一结果大大提高了数据处理的效率。

(1) To measure a deviation of objective surface of machine parts, a new interferometer by using the CCD image sensor instead of hologram in the previous proto-type interferometer was developed. The Introduction of the CCD image sensor gives stable measurement.(2) The brightness data of each interference fringe pattern to reference and objective surfaces is converted to discontinuous phase difference. Also a proposal in this report, that is, the computer holographic technique, was applied to the data processing. The technique enable phase to concatenate efficiently.(3) The interference fringe pattern on the CCD image sensor involves a thick density of fringes, and it also contains a considerable amount of noise. The problem to be solved such as pattern connected between neighboring fringes is remained, even if it is relieved by the technique mentioned above. For this problem, some methods by reducing the fringes by inserting concave lens in the path of the reference beam, by applying target surface form and by the heterodyne method using two beam sources with different wave lengths were introduced in the experiment. These methods were successful for data processing from the brightness data to the form deviation.(4) It is not distinct how many interference fringe patterns should be taken by phase shifting method to calculate phase difference from brightness data of interference fringe pattern. From a series of experiment using some plane test pieces, it is cleared that 6-8 sheets of the pattern are sufficient. This result increases considerably efficiency of data processing.