Dear DAVID Community,

As you are probably aware, the entire DAVID company, team and product range has been integrated into HP Inc.
This forum is now read-only. Relevant content has been migrated to HP and merged into HP's Support Forums on November 1, 2016.

To start new discussions for 3D Scanning, please register and post your new topic at the HP Support Forums

High Accuracy Scan Using SLS

Problems, solutions and discussion about scanning with a video projector and stripe patterns.

High Accuracy Scan Using SLS

Postby bacaglar » Thu Jul 24, 2014 9:12 am

Hi everyone,

I am planning to use DAVID for measuring the thickness of planar surfaces (where thickness change is less than 10 mm) and the surface has dimensions of 600 mm x 300 mm. The thing is I am not sure if I can get the thickness for this size of a surface accurately. I know from some papers that it is possible to get accurate readings using structured light but I want the accuracy to be less than 0.05 mm on such a big area. I have done some preliminary work by scanning smaller areas to see how accurate results I get.

Below is a typical measurement I get. In this measurement I placed an object (which is flat) on a flat table and performed the scan. Red dots show the object I placed and how do I get these red dots? I use the RANSAC method to find a plane with a tolerance of 0.05 mm and those red dots are the inliers (within the 0.05 mm tolerance). Blue dots are the outliers and it is normal because I know those belong to the table. However things get interesting when I fit a plane for table using the same method (fit a plane for the blue region).

Image

This time, there are some outliers on the table surface (especially on the right side of the figure) even though I expect them to be inliers. Things get even more interesting if the distance between object and camera increases (distance between object and the projector increases as well).

Image

Below is the RANCAS fits for the object and the table. Some outliers are introduced in the object plane and half of the table is outlier of table plane.

Image

Image
There are many more scans like this, but I won’t put all of them here. My observation was if I scan a larger area the quality of the scan decreases. Also most of the time, the outliers are on the edge of the figure, so I am putting the blame on the lens distortion.


These scans are performed using Canon EOS 550D camera with 50 mm f/1.8 lens. At first, I was using 18-55 mm Canon lens and results were even worse. I read somewhere, that Canon 50 mm f/2.5 Compact Macro lens has much smaller distortion than my current lens so I am planning to switch to that lens (once I am convinced that I can go on with structured light scanning). However, I am using the Canon camera at 1056x704 resolution and I think that the low resolution may be contributing to quality of scans as well.

In these scans I have used Casio XJ A-140 DLP projector. I also borrowed two LCD projectors (Epson EMP-710 and Panasonic ……) and those projectors performed more or less the same with DLP one. So I am totally lost projector-wise.

By the way, in all of my scans I had a horizontal triangulation angle between 25 and 30 degrees. I focused the projector and the camera on the object. I can share any other detail of my scan if needed.

So, I would appreciate if anyone has any suggestion in which direction I should go? Should I buy new hardware, or there may be some detail I am missing in calibration and/or in software settings. I did not spend any money until now, I have been going on by borrowing stuff from people and I want to spend the money wisely and make sure that buying new hardware will worth it.

Baris
bacaglar
 
Posts: 1
Joined: Tue Jul 22, 2014 7:20 am

Return to Structured Light Scanning

Who is online

Users browsing this forum: No registered users and 1 guest

cron