In the period 1987 to 1988 Mauritz Aas did his cand. scient. thesis in the Cooperative Phenomena Program. Aas et al. worked with an environment on an IBM AT computer for capturing pictures with a video camera. They started out with a PC Vision Frame Grabber Adapter and put a large effort in to minimizing the time used for capturing the frames. Their system had a resolution of pixels with 8-bits grey level resolution for each pixel. The time necessary for refreshing the video camera, grab one frame and save it to the hard disk with this system was 2.1 seconds. This was almost as low as the theoretical limit. Together with software for controlling the frame grabbing they also developed other simple image processing tools.
An attempt to use this system together with a JVC Model GX-N7E color video camera and a JVC VCR was made. The system was only used to map one structure. As explained in the introduction to this chapter the contrast of the system was not good enough for this purpose. The CCD-chip in a video camera is continuously exposed to light. It works as a digital scanner and is continuously read out line by line and reset. The read out time is very short so the chip is not able to accumulate much charge between each read out. Video cameras work very well under good light conditions, but even professional television cameras do not work well under bad light conditions. (In all television studios there are lots of spot-lights for improving the light conditions.)
The system also had stability problems. The reason for this could be that we used a different video camera than the system was optimized for. Another reason could just be bad connections because the system had not been used for a couple of years. However, we did not investigate this any further.
The last problem was that the frame grabber card was not designed to give square pixels. The pixels had a different horizontal and vertical size. They were only 2/3 as wide as they were high. The images were also stored in a special block format to minimize the storage time. To be able to do calculation on these images they had to be transferred to an Apollo work station. They also had to be transformed to the standard format used at the Cooperative Phenomena Group. In this format the pictures are stored as two lines. The first line contains information about the picture, and the second line contains the picture stored binary. The information line has three numbers separated with spaces and a text string which gives information about what kind of device created the picture. The first number tells the resolution of each pixel. This should be the number of bits minus one. For 8-bits resolution this means that the first number should be seven. The two following numbers are the number of pixels in the y- and x-direction, in that order. The pixel data are stored binary in x-directional order starting with the left pixel in the upper row and ending with the right pixel in the last row.
A filter program which converted from the PC Vision standard format to the Cooperative Phenomena format was written. The filter also stretched the image so that the pixels got a square size. This was done by putting in an extra pixel for every second horizontal one. The value of the extra pixel was the average of the ones to the left and right.
A picture of the PC Vision block format and a converted picture are shown in pcvisionformat. A description of the filter program can be found in appendix .