Panorama

[FrontPage Include Component] Previous
Home
Up
Next

NAVIGATOR: Back - Home > Adi > Services > Support > Manuals > Apas > Dos :

Home

Panorama

Up
Kinetics
Analog
Delta
Gait Analysis
Anthro
Rotations
Panorama
DIGI4

Categories

Products
Services
Site Map
Search
About us

Search

Contact us

General info

+1 949 858 4216

Sales & support

+1 619 992 3089

E-mail

Information

Sales

Support

CHECK OUT

[FrontPage Include Component]

Ariel "Panorama" Panning Head Operation


INTRODUCTION   The Direct Linear Transformation (DLT) is a technique commonly used to locate spatial points filmed with two or more stationary cameras. The DLT does not necessarily yield coefficients which correspond to an orthogonal orientation matrix of the image to object coordinate system. It has been demonstrated that increase accuracy for the reconstruction of points is achieved by imposing an orthogonality constraint upon the optimization procedure used to obtain the DLT coefficients. This DLT with orthogonality constraint is called the modified DLT (MDLT).

Ariel Dynamics, Inc. has developed a technique called physical parameter transformation (PPT) which allows the use of panning cameras. The PPT is built upon the colinearity photogrammetric relations from which the DLT is derived. Like the MDLT, PPT is implemented such that orthogonality of the orientation matrix of the image to object coordinate system is guaranteed. PPT with panning will be demonstrated to have greater accuracy than the DLT.

PANORAMA is an option of the Ariel Performance Analysis System that allows spatial coordinate reconstruction from panoramic or still camera views in any combination. PANORAMA consists of the panning angle measurement hardware and sophisticated software for data transformation after the digitizing process in the APAS.

EQUIPMENT SET-UP

  1. Mount the Panning Head to your tripod. Be certain that the screw on the tripod threads into the threaded hole in the tripod mounting plate of the Panning Head. Screw tightly to ensure a good connection.
  2. Disconnect the upper half of the Quick Release Mechanism by pushing the button on the end of the lever, and simultaneously pushing the lever to the side. The upper shoe half will pop off. Attach the shoe half to the base of the camcorder and attach it to the camera. The upper half mounts to the camcorder. Be certain the attachment is secured tightly. Lock the camera to the Quick Release Mechanism of the Panning Head.
  3. Disconnect the viewfinder cable of the camera. Attach the Panning Head cable with the two way connector to the camera EVF/CG terminal jack. Finally, reattach the viewfinder cable to the two way connector. Turn on the camera power switch.
  4. Once the viewfinder cable has been connected to the two way connector, the Power Indicator will light up, and then a white horizontal bar will appear in the viewfinder. To check if the horizontal bar functions correctly, press the Reset Push Button and the right edge of the bar will be set at the middle of the viewfinder's width. The right edge of the white bar should move proportionately to the camera angular displacement.

DATA COLLECTION

FILMING AN ACTIVITY   These rules must be followed to ensure that subsequent analysis will be correct and accurate:

bulletAt least two cameras must be used for three-dimensional analysis. The number of cameras that can be used is unlimited.
bulletWhen filming an activity, it is essential that all cameras record the event simultaneously and include a synchronizing event..
bulletIf stationary cameras are used, they must be placed so that each one is able to film the entire sequence of motion without camera movement.
bulletIf a combination of panning and still cameras are used, it is recommended to use panning cameras for recording side or wide views, whereas still cameras are better to use for front or back views.
Filming A Panning Camera View:  The bar range should be selected using switches 1 & 2 on the Panning Head and set according to the area in which the activity will be recorded. In order to increase the accuracy of the activity's spatial reconstruction, it is recommended to select the lowest possible range. However, the operator should be sure that the range he or she selects is sufficient to cover the whole activity.

Bar position in the image and bar width should be selected by selecting the Bar Width switches (3 & 4) and the Bar Position Switches (5 & 6) located on the rear panel of the Ariel Panning Head.

CALIBRATION OF THE PANNING CAMERA  For the panning views, it is recommended to position Two Calibration Objects; one at the left side of the panning range, the other at the right side of the panning range. There should be a minimum of 6 control points on each calibration object with known X, Y, and Z coordinates from a single reference point.

In order to have the calibration objects proportional to the bar scale, the camera should be directed to the center of the panning angle and then push the Reset Button on the panning head.

Then film the left and right calibration objects. It is recommended that the control points be filmed before and after the activity is recorded. However, one frame for calibration is often enough if the user is certain about the control points in the procedure.

NOTE:
The Reset Button should not be pushed between the filming of the calibration objects and the filming of the activity to be analyzed.
GRABBING A PANNING VIEW  Frame grabbing for Panning Views is identical to the grabbing procedure for "fixed" APAS camera views with one additional recommendation. The left and right calibration views should be grabbed and stored in a file separate from the panning data file. This would mean that each "panning view" would have two .VID files stored on the E:\ disk. During the Digitizing process, the F1 and F2 keys on the Control Point Menu can be used to select the grabbed calibration file.

DIGITIZING A PANNING VIEW  As with normal Ariel Performance Analysis System (APAS) sequences, data collection begins with the videotaping process. Most sequences can be recorded in their entirety with one side panning camera and one front/back stationary camera. The panning angle is provided by an optical encoder yielding 10 minutes of resolution. Panning axis location and direction were determined by performing 2 normal PPT calibrations corresponding to different panning angles. The two calibrations yield positions of the camera which only differ by a rotational displacement about the panning axis; thus, the calibrations provide enough information to determine panning axis location and direction. Since the displacement may be expressed as a function of panning angle, axis direction, and axis location, the PPT coefficients corresponding to any panning angle about a single axis is determinable.

DIGITIZING  Digitizing is the process of converting film or video image sequences to computer image sequences (stick figures). To produce computer images from film or video images one must convert the recorded patterns to numbers or digits, hence the term digitizing.

Each frame of film or video is a picture and contains a great deal of information. Usually the information is far more than is needed to analyze the motion of the subject. Therefore, the digitizing process becomes one of digitizing the subject rather than the entire frame. Human subjects consist of various body segment (limbs) connected at body joints. If the location and identity of each of the body joints are known then the location and orientation of each of the body segments (therefore the complete subject) can be easily determined. The APAS takes this approach to the digitizing and saving of images of subject motion; namely the determination of joint location for each body joint in each frame of the film or video recording.

Digitizing a panning view is exactly the same as with stationary cameras up to the point of entering the View Information. The Ariel Digitizing module has been enhanced to allow for use of the panning camera using the custom panning head. The PAN option should be toggled to YES to indicate a panning camera. When the PAN option is turned on (YES) and F10 is pressed to proceed, you will then be required to enter the Panning Options.

  • "Info" is selected to enter the panning camera information. View options includes a new entry "PAN" with 3 sub-menu items, Info/Right/Left. Info allows the user to specify a string & #panning/control points for each of Right/Left panning calibration. Right/Left allow the entry of the XYZ coordinates for the Right/Left panning/control points. When this option is selected, you will proceed to the Panning Calibration Information menu.
  • "Right" and "Left" are selected to enter the control point locations for the Right and Left calibration fixture. With PANNING views it is necessary to digitize a set of pan/control points with the camera panned RIGHT & a set with the camera panned left. Both Right/Left sets must be digitized to process the data. The XYZ & UV values for a panning view are stored in a separate file. Each panning view has an associated file with extension ".xC". Eg: TRACE.1T & TRACE.1C
  • "Quit" is selected when the user has completed entering the Information for the panning options.
  • "F1-Read" is selected to read the digitized values from another file. The Panning Calibration Information Menu provides the following options:
    bullet"PanRange" is used to set the panning camera range. This is determined by the DIP switches #5 and #6 on the panning head hardware. Select the appropriate range using the up/down arrow keys.
    bullet"#Points[Right & Left]" is used to specify the number of control points in the left and right control fixtures for the panning camera.
    bullet"ID[Right & Left]" is used to specify an identification label with the corresponding control points.
    bullet"F10-Done" is selected to proceed with the digitizing process.
    Digitizing a panning view is very similar to the stationary camera view with the following exceptions:
    1. When digitizing a "panning" view, the "fixed point" is replaced by the "Pan Bar". The panning algorithm uses the panning bar length to calculate a DLT set separately for each frame.
    2. Digitizing Control points. For fixed as well as panning cameras, the program allows the user to specify a different GRAB file from the one used for data. The user can also select any image# within the file for use with control points. Use "F1" to select a GRAB file. Use "F2" to select an image# within the file. Use "F3" to toggle between Right/Left [indicator in upper/left corner of Select Control Point Option menu box.
    TRANSFORMATION  The Transformation process allows for the PPT algorithm explained earlier. Any view defined as panning uses PPT in the panning algorithm. Panning views automatically use PPT. This specification affects only FIXED camera views. For any view defined as "panning" the algorithm will use Right/Left control point data to calculate separate DLT sets for each frame.

    When "Consistency Tests" is selected, a different report is generated depending on whether view is DLT, PPT or PAN. "Residuals" are always reported in digitizer units with a range 0->9000 full range. Since the screen is of resolution 640x480 one converts to pixels as follows:

    		1 "x" pixel = 9000/640= 14 digitizer units
    		1 "y" pixel = 9000/480= 19 digitizer units
    
    1. DLT- As before. Residual is given in Digitizer units. For each point [u,v] digitized data is compared to where it is expected to be given the best fitted camera parameters & accumulated for all points. Residual as well as Residual per degree of freedom is reported. The following define mathematically the quantities:
      			ChiSq= SUM[(u-u0)**2 +(v-v0)**2)
      	    	 	Residual=Sqrt(ChiSq)
      			ChiSq/DOF=ChiSq/(2*#CTLPTS-11)
      			Residual/DOF=Sqrt(ChiSq/DOF)
      
      			where [u,v] is observed image location [0->9000 range]
      			& [u0,v0] is the expected image location
      
      The #degrees of freedom [DOF] when fitting parameters to data is the # of measurements [2*#CTLPTS] minus the # parameters in the fit [11].
    2. PPT- DLT parameters corresponding to the PPT are reported along with the PPT parameters. The PPT parameters are defined as follows:
      		A1/A2/A3-	Successive rotations about z/x/y needed to
      				specify camera orientation in lab frame
      		R1/R2/R3-	[x,y,z] Location of camera projection point
      				in the lab frame.
      		g1/g2-		Projection plane origin in digitizer units
      		h1/h2-		distance from projection point to projection
      				plane in digitizer units [u/v separately].
      
    3. Pan- PPT parameters for Pan Right calibration and panning info for best fit panning to left camera calibration.
      		A1/A2/A3-	Successive rotations, A1/A2 about y/z to locate
      				panning axis, then A3 about Y [pan angle].
      		R1/R2/R3-	Pan Axis location in lab.
      
      Residual is reported for pan left calculated directly from pan left calibration data and transformation using right pan calibration transformed according to "pan" solution. "Direct" will always be smaller [better]. The closeness of the two measurements indicates the "goodness" of the panning solution.
     

    Previous Up Next

    This page was last modified on 12/07/2008 at 22:47 PST. Copyright � 1994 - 2002, all rights reserved, Ariel Dynamics Inc. Please send your comments or feedback to webmaster@arielnet.com or proceed to our feedback form. This page has been accessed many times since Dec 12, 2002. Our privacy policy is here.