Week 6
To begin the 6th week we started to combine the threshold code with the SURF code more. The code is now able to search for a specific color and image that is given to it. We have the program working well with a few bugs where it crashes every so often. Our main goal is to move the equipment from the ERF lab to the LSRI lab so that we may begin testing with the overhead camera. This should allow better detection for the SURF program to run off of while there is better lighting and a clear distance between the camera and object. We walked over the equipment to the LSRI lab to discover we had no table for the computer to sit on. So we came back to the ERF lab so that I could continue transcribing.
We eventually made it back to the LSRI lab so that we could begin testing the program. We set up the overhead camera by taping it to a pole and making sure it was parallel to the floor. The original camera had something wrong with it, so we couldn't use that one and had to use a wide-angle lens camera. With the wide angle lens the camera has a bigger surface area to look at. It was my job to test which square sizes could be picked up and what the range was for the sizes. Tia had already finished a chart so that I could fill in the information easily.
I began testing the pieces and at first it was going smoothly. The two larger squares worked just fine, the 1.75 and 1.5 inch squares. Anything after that became a little fuzzy. Looking on the screen I could tell the camera was having problems picking up the pieces, so I started to tamper with the lighting of the area, brightening and dimming the lights to see if it would help the problem. No matter what I did it still seemed to have trouble, so I looked closer at the screen to see if it was an immediate problem. It looked as if the table was reflecting the lights weird. So, I then tried to use something smoother and less reflective to cover the table. The only thing I could find in my workspace was a large piece of cardboard, and I believe the squares were too close in color to be discriminated by the camera. I tried moving the squares in several different areas and even in just the plain middle of the table to see if it would get picked up, but nothing. I decided that since I couldn't find the others I might as well try finding the exact points for the first two that were found.
This week was short because it was the Fourth of July, so we were only in for three days. I felt as if there was a decent amount of work completed, but there is still much more to be done.