Scan_Calculator_linear

Author: Peter Stierlen (peter.stierlen@t-online.de)

History: 2009/04/27 (v1.00 beta): first public version

2009/05/07 (v1.01): “spacetime” calculation removed

2009/05/11 (v1.02): improved “standard” calculation with better

skeletonization

Download: Scan_Calculator_linear.jar

Installation: Download Scan_Calculator_linear.jar to the plugins folder and restart Imagej. You need the intrinsic camera parameters of your cam system if you want to use the undistort function.

Description: This plugin allows you to calculate 3D data out of 2D scanline data captured with a linear laserscanner. This plugin can be used for linear scanners with camera and line laser fixed to each other moving together linear to the object. Different scanner configurations can be used very easily.

Limitations: Different to the plugin Scan_Calculator the spacetime analysis is no longer enabled in this one. Spacetime analysis as described in the listed paper is working for scanner setups with fixed camera/object positions.

Data-Type	Format	Extension
Polylines	Rhino3D - polylines	*.txt
Points	Rhino3D - points	*.txt
	x,y,z	*.txt
	x_y_z	*.txt
Meshes	STL- meshes	*.stl
	PLY-meshes	*.ply
	OBJ- wavefront meshes	*.obj

Calculation of 3D data from 2D image stacks captured by a laser scanner

Open your original distort stack

Start the plugin Scan_Calculator_linear

Clicking the plugin Scan_Calculator_linear you will start up in the main menu which is split in two different sections.

In the upper section you can check and edit the basic scanner setup data. In the lower section you can find all the functions needed to control the data calculation process. If you do not want to undistort your 2D image data you can directly start the calculation from the main menu. But I will now describe the procedure for the complete plugin configuration.

!!! You must input the linear Step Pitch value. This value describes the movement of the setup between two captured images!!!

Scan_Calculator_linear – Maim menu

Change to the Camera Prefs Menu

Camera Prefs – parameters for undistortion of images (zhang calibration parameters)

In the Camera Prefs Menu you can check and edit all the camera parameters needed for the undistortion of the 2D images. Since Version 1.00 beta Focal length, Chip width and Chip height are no longer needed for the undistort process.

You can also load the calibrations data directly from the zhang output file Params.txt

by File>Load Camera Prefs>…\Params.txt

The data will directly be written into the input-fields.

Load Camera Prefs – import the zhang calibration parameters for the used camera

Please keep in mind that the camera parameters are only valid for the calibrated camera and camera configuration (zoom).

Definition of 3D bounding box

Change to the menu 3D Bounding Box. Only the points included in the bounding box will be processed. It will be a helpful option to minimize the number of “nirvana” points.

Definition of the size of the 3D bounding box

The height of the bounding box is defined similar in the z-direction. Half of the box is z=height above the xy-plane and the other half below. In most cases it will be adequate to define the frontside and the backside of the 3D bounding box to reduce the excessive points.

3D Calculation

Undistort

Please keep in mind that grabbed pictures have a optical deformation which can decrease the accuracy of the 3D data significantly. You can undistort the image stack directly by clicking the Undistort button.

The undistortion process is not very quick … wait.

Save the resulting stack as a Tiff-Stack. Keep in mind that the metadata from the source stack will not be stored in the new stack file. If you want to read the metadata including the scanner configuration data you must also open the source file and read the meta data from this one.

Read metadata

If you captured and saved the 2D image stack with my Scan_Capturing plugin you can read in the scanner configuration data by clicking the Read Meta button directly from the stack. You can check and change the data in the upper section of the main menu if necessary.

Make 3D

You can choose 5 different types of calculation. Naturally the trigonometry is still the same. You can find the differences in how the scanline in the 2D images will be skeletonized.

Standard (Gauss):

Each slice in the stack is processed separately. In each slice the program looks line by line for the position representing the center of a Gaussian intensity distribution. The center position is calculated in “sub-pixel” accuracy and will be used as is in the following calculations of the 3D data.

Standard (Fourier):

Each slice in the stack is processed separately. In each slice the program looks line by line for the position representing the center of the intensity distribution by using a Fourier method. The method is described in http://www.numatec.de/Fachaufsatz.pdf (German language) but you can also download a excel sheet demonstrating the simple method http://www.numatec.de/FourierPos.zip

The center position is also calculated in “sub-pixel” accuracy and will be used as is in the following calculations of the 3D data.

Quick:

If you have a stack with skeletons you can directly process the 3D data with this option. Please keep in mind that in opposite to the other methods the accuracy is not “sub-pixel”

MAX:

This type looks for the most “black”-pixel in each line. If you have more pixels with the same value it will take the most right one.

Saving 3D Data

The calculated 3D data can be saved in different point or mesh formats. All formats are ASCII –based.

Data-Type	Format	Extension
Polylines	Rhino3D - polylines	*.txt
Points	Rhino3D - points	*.txt
	x,y,z	*.txt
	x_y_z	*.txt
Meshes	STL- meshes	*.stl
	PLY-meshes	*.ply
	OBJ- wavefront meshes	*.obj

Post-Processing of 3D data

In most cases the 3D data require a little bit of post-processing before you can use them in other CAD, CAM or 3D-visualization programs.

I advise you the free software Meshlab http://meshlab.sourceforge.net/

Fantastic tool!

Happy meshing!

You are visitor

counter