Index
Home
Next: Display Adapters
5 Camera Module Construction
Section 3 described the camera used in the project, as well as how to modify them and how to mount them to the base. Then section 4 described the sync circuit board. Much of the information included here has been provided in previous sections, but the intent is to tie everything together into a module that can be connected to and operated. The focus of this section will be the base, the assembly of the cameras and circuit board onto it, and operation.
The camera module consists of two board cameras, a printed circuit board that holds the synchronization and other control circuitry, and a mounting base. The base is composed of a 3" by 6" piece of 1/8" thick clear Lexan.
Each camera is mounted in a 2 1/8" long narrow channel by #4-40 machine screws, nuts and washers; each channel is long enough to move the two cameras together or apart. Spacing between cameras can range from about two to four inches. An adult's eyes are typically spaced about two and a half inches. Moving the cameras further apart will enhance the stereoscopic effect, but spacing them greater than about three and a half inches can cause eyestrain, especially when viewing close-up objects. In fact, testing in "close quarters" had to be done with separation at its minimum of about 2". Trying to view objects closer than about six feet caused a great deal of eyestrain. Because the disparity between the two view was so great, it felt like having to cross your eyes to get an object into view. (Try to focus on a finger touching your nose.)
Each mounting channel is about 3/16" wide. This was done to allow the camera to be rotated a few degrees within the channel, so that the line of sight from the cameras will converge at a fixed distance. However, experimentation showed that this has no benefit. The cameras should be aligned on either the front edge of the channel or the rear edge. There is no accommodation for focusing the cameras beyond the lens provided with the cameras.
The two cameras should be at the same vertical height and vertical angle. It is much easier to make them achieve the same angle, whether greater or less than ninety degrees, than to make them exactly vertical. Vertical height will depend on tolerances of the camera lens housings (out of our control), tolerances of the mounting hardware, and how the cameras are mounted. Section 3 explains how to use cellophane tape to elevate the front (or rear) edge of one camera to make both cameras the same approximately vertical angle. Once the tape has been applied, that camera shouldn't be moved - all camera separation adjustments should be done with the "other" camera - the one not leveled with the tape. Not moving the leveled camera is necessary in order to prevent the tape from being wrinkled or pulled up.
The cameras connect to the circuit board by way of up to six wires from each camera. (See Figure 16.) The left camera connection will require six wires, and the right camera requires only five wires. A Molex wire-to-board connection scheme is specified in the Bill of Materials (Table 2), consisting of connector housings and crimp pins. However, the wires can also be soldered directly to the circuit board.
Three connectors at the rear of the circuit board provide power input, video output, and shutter glass control signal output. There is also the potentiometer that is used to adjust the timing of the shutter control signal.
The circuit board is mounted using #4-40 screws, nuts and washers. A #4-40 nut is also used as a spacer to hold the circuit board above the base to accommodate IC and connector pins and component leads.
The base rests on six clear rubber feet, one in each corner, and one each centered on the front and rear edges.
Figure 33 Camera Module Base Dimensions
Figure 34 Camera Base
Figure 35 Camera Base with Sync PCB
Figure 36 Camera Base with Sync PCB and Cameras
The module is pretty basic, and there are only two things that can be operated (other than the on-off switch): camera separation, and shutter control.
The cameras are mounted in the channels in the base. The nuts and screws that hold the cameras to the base are loosened and each camera is moved laterally in its channel.
The distance between the two cameras determines the magnitude of the stereo effect at any specific distance. In his book The World of 3-D, Jac. G. Ferwerda explains that a "standard" separation of 65mm (about 2 1/2" - the average distance between an adult's eyes) will provide a good stereoscopic effect for distances of about six feet to about 200 feet. Therefore, separation should be reduced to minimum for viewing distances of less than six feet. The shortest separation that is possible with the carved-into-Lexan dimensions described above is about 2 inches.
Here's how to determine which camera is being enabled for each eye. While wearing the shutter glasses and viewing stereo video on a TV or VGA monitor, close your right eye and wiggle your finger in front of the right camera, around two or three inches away. If you can see your finger wiggling, then the correct video field is being enabled for each eye. If you don't see a finger, then R3 (the pot) must be adjusted. As you turn the pot, you will see a bar going up or down the screen. Move this bar to the top (or bottom) of the screen. If the bar is moved to a position opposite of its original location, the video enabled for each eye will be reversed.
So if the bar is at the top and you can't see your finger wiggling in front of the left camera while your right eye is shut, move the pot so the bar moves down the screen. If you move the pot too far clockwise the glasses fail to alternate, and both lenses become transparent. This happens because the pot was adjusted to a value that keeps the timer from expiring, and the shutter control signal has become a DC level.
If the camera module is in a remote location, then obviously the previous paragraph won't work. Then the technique will be to view the scene with each eye, one eye at a time. When the shutters are displaying the correct view to each eye, each eye will see the same scene, but at a slightly different angle. Let's say we're looking at a book, face on. If the left view shows the front and a bit of the spine on the left side of the book, the right view will also show the front, but will be more square on. That is, the spine that was seen in the left view may be much less visible in the right view, and possibly not visible at all. Another way of saying it is objects in the left view will be rotated slightly to the right, and the same objects in the right view will be rotated slightly to the left.
Index
Home
Next: Display Adapters