Progress Reports for Andy

   
 
April 3, 2002

Last week I started the redesign of the image processing. I ended up incorporating a couple of different ideas. First I used a 5 deep array that keeps data for a 5 degree angle in the picture. The goal is to find which section has the highest concentration of color closest to the reference color. This has a couple of problems with it that I am still working out.  One advantage is that different lighting levels should have a minimal impact on the image. But this also means that colors that are a darker shade of the target are considered the target. There is also the problem of a lot of 'noisy' values. Hopefully this week and this weekend I can iron out these problems. It is questionable as to whether the image processing will work well enough to connect to the car even if I can get the image processing to 'work'.

March 27, 2002

This week I was all ready to come back and get the senior project moving.  Alas nothing about this senior project goes the way I plan. I found a simple error that had caused the image processing to fail. Once I solved the problem I found another major flaw. Since the camera is not very good (no digital webcams are) the entire image is very bright which decreases the intensity of the red object I am detecting. Then in addition because of this brightening brown objects appear more red than they should to the camera. This causes the image processing to fail. So I am trying to think of ways to salvage the image processing. The next thing I am going to try is to make the column vector a 5 element deep array. Then separate the intensity of the red pixels into 5 levels. This will hopefully pull out the areas with the largest concentration of high intensity red pixels I hope.  This is about the only solution I can come up with to separate the red pixels and determine which ones make up the red object and which ones are just slightly red background noise. That is this week's project along with about 3 other tests :-(.

March 13, 2002

This week was a smashing success (for once). This week I connected all my modules together. I found the problem in my twain driver and once that came together the rest of the modules fell into place. I was able to finalize a couple of classes. I did find a slight problem with pointer manipulation for the image processing that I hope to have resolved the first week after break. Depending on how the image processing works after I fix the pointer problem I might need to work on redoing the image processing. In addition, I found out that the capture of an image takes 4 seconds. I will be working on threading the application more than likely. I will also try professor Malinowski's suggestion that I use a static array for the image passing.  These will all be addressed right after spring break. After spending about 20 hours on the project the past couple of days I am ready for break.

March 6, 2002

This week I worked on the class to create the command signal for the microcontroller. In addition, I have started to connect all the classes together. As soon as the TWAIN driver works I will have a working program that is ready to test. I hope to finish it up this week before all my tests at the end of next week.

February 27, 2002

This week did not see any great progress like the previous few. I investigated a couple of laptops because of the pressing need of getting one for testing purposes. I also tried working on my senior project tonight but discovered that all the work I performed last week has disappeared. So I have to start back with the first iteration of the image processing. In addition I have been thinking of a couple of improvements off the bat to the first image processing algorithms. This week I will try to finish the serial port driver and also finish the first iteration image processing.  Then next week will hopefully be the week I finish all but the twain interface.

February 21, 2002

This week I tried to finish the driver for the twain interface (again). After banging my head into the wall most of the day Sunday, I gave up on it for a few days to collect my thoughts. Last Thursday I wrote the first iteration of the image processing but had a couple of errors I have been debugging. This week I will be taking measurements on the accuracy of the image processing algorithm.

February 13, 2002

This week hardly any work got done on the project because I was gone the entire weekend and Monday, which meant I had to catch-up on homework on Tuesday. So no software was written this week. I have started in on the theory behind the image processing for the first iteration. This is all laid out (a rough idea basically) in my notebook which I am turning in this week. This next week will hopefully have a breakthrough with the TWAIN driven and the initial image processing algorithms.

February 2, 2002

This week was a very slow week for the project. After spending all day last Thursday developing code that still has a bug in it, I went on a search for a camera that would satisfy our needs. The camera I originally bought and any other camera I have seen do not support twain properly. I finally found one on Tuesday, Logitech WebCam Pro 3000. So I am back to writing the twain driver. I expect to have it completed by next week. This may be a problem though because I will not be on campus fri-mon. Total time spent the last week was about 12 hours. I am also entertaining the idea of using an artificial neural network to do some of the image processing because it would speed up the process.

January 23, 2002

This last week was marked with disaster.  The camera (DVC325) I bought doesn't support disabling of the user interface so my TWAIN driver was halted on Saturday.  Now I am rewriting the driver to make it more easily read and expanded.  In addition I am composing a class that will interface with the TWAIN driver and update the capabilities to work as directed.  The next step is to write a class that will implement an entire session for the program.  This will be the wrapper class.  I have already dedicated over 40 hours since school started and don't see the end in sight.  I will upload the code for the TWAIN driver when I am ready to release the first version, this will hopefully happen next week.  In addition I am uploading a graphical approach to the classes under code.  This is in addition to updating the website.

-- Current work looks like this. :-)

 

January 27, 2002