Category: Logan’s Status Reports

This week I worked on the Final Report as well as prepared for the Final Presentation which I am responsible for. Next week I will continue to work on the report, and either assist my teammates integrating the Jetson or on other tasks such as graphics depending on our mitigation plans.

During the start of the week I  continued to test the CV with video to see what kind of performance we could get at full framerate. Since we were not able to power the Jetson, I used my laptop. While it doesn’t have the same specs as the Jetson, it would appear that computationally, we should be able to use at least half of the maximum framerate without latency issues. The next step is to test whether our transmission capabilities will also allow this, and then see how the Jetson performs.

I also requested the power supply part for the Jetson.

This week I was able to get the CV portion of our MVP working on some test images, and start testing the video input of the CV code. The performance on video input (given the numerous frames) is potentially an issue, a possible solution is to ignore most of the frames, only looking at every third (for example) frame. This should still give us enough resolution to get information on the swing while cutting down on the data we need to process.

 

This week I continued to work on the CV with our new strategy of color recognition as opposed to the previous strategy of contours. Around the middle of the week I ran into some issues but I reached out to Professor Savvides for help and he gave me some tips about hue/HSV and pointed me to some resources which continue to be helpful.

Despite the unforeseen challenges this portion of the MVP should be done by Monday so we can proceed with integrating next week.

This week I continued with the CV portion of our project and had success on detecting position in most cases, but we ran into considerations about how the cameras would recognize the paddle in different orientations. We decided to change our strategy from contours to color recognition which would allow more flexibility in recognition if the paddle was closer to parallel with the floor. This strategy would also open up the option of coloring the edge of the paddle so that could be recognized as well.

This week I will go forward with this path and hopefully finish it up so I can turn my attention to more nuanced issues with the CV or assist on other subsystems.

 

I did a very small amount of work on CV this week, as I was severely ill from the middle of the day Monday until Thursday evening. I am posting early as I am unavailable for the next 2 days due to other commitments.

Clearly, I was not nearly able to meet schedule for this week but I plan to make up for this some time next week and over break.

This week I continued to look into CV while awaiting materials. As I further delve into the CV content, I have increasing doubts that we will be able to meet latency requirements. After discussions with the team we have developed fallback plans to adjust the speeds of the ball to buy more computation time, either by scaling all speeds globally, or locally slowing down opponents’ shots while keeping the player’s shots in real-time. We have also considered scaling up the size of the court to make the game more similar to tennis. I also began preparing for my assigned design presentation.

This week’s progress was decent, clearing up details on our design and creating mitigation planning for the foreseeable latency issues we may have. Now that our materials have arrived we can really make progress on our subsystems .

This week I worked on tinkering with OpenCV to get oriented with the computer vision aspect of our project. Specifically, tutorials on the core functionality, image processing, camera calibration, and 3D reconstruction along with some other documentation.

Next week I plan to get some initial metrics on the accuracy and latency of the software. Ideally, we will have our cameras for this but I should be able to do some cursory testing regardless.  This will also give us an early chance to develop workaround or solutions for high latency which is a major concern at the moment.