Previous | Next | Brown & Sharpe
Topic: Cone measurement
Conf: Brown & Sharpe, Msg: 1745
From: Deleted User
Date: 4/8/2000 11:37 PM

Two of the geometric shapes PC-DMIS has problems with sometimes is short and fat cones/cylinders. If you take the basic number of hits (6), for some reason it sometimes creates a cone/cylinder roughly 45 degrees from what it should be. What appears to be happening is the algo is taking two of the top row hits and one of the bottom, and two of the bottom and one of the top in the construction. The best way I found around this problem is giving the algo more data, hence, take a lot more hits. This will force the algo to construct the feature only one way. Remember, cones are constructed using a best guess, and with only 6 hits, its doing the best it can.