逐次多点法による表面プロフィール計測方法

計測時間の短縮

抄録

We'v already proposed Serial Multi-Point Method (SMPM) for surface profile measurement.
This method is very powerful because the influences of the sensors head's fluctuations can be canceled completely and the accuracy is improved with only increasing the number of sensors.
But, the amount of calculation, calculation to make a pseudo inverse matrix, increases in accordance with the increases of sensors and measuring points. And it resulted in a longer measuring time.
In order to overcome such a weak point of SMPM, we'v invented a supplemental method to reduce the amount of calculation.
In so doing, we'v checked and eliminated needless calculations, and found a special characteristics of the square matrix to be inverted, characteristics specific to SMPM.
So, we'v established a recursive method to make an inverse matrix without using sweeping-out method upon a matrix of large size.
This recursive method is very effective to reduce the amount of calculation and to shorten a measuring time.
As the result, with our method newly proposed in this report, the amount of calculation become less than one tenth or one hundredth.
So, SMPM became more applicable to the time-strict purposes even if the number of measuring points varies diversely and we can't prepare the pseudo inverse matrix in advance.