Motion and Speed Detection

Our motion detection code works to detect when there is motion in any one of three side by side sections of the camera's view. Today, I worked on sending values to Serial for when each of the three sections detects any motion. I also began work on speed detection. One challenge we faced regarding speed detection is that we have no way of knowing how far away a person interacting with our sculpture is from the sculpture. This means that it is difficult to to gain an accurate reading on speed detection. For now, we are going to assume that a person interacting with our sculpture is 5 feet away from it. This should work relatively well for detecting relative speed of motion from different distances.

- Vicky

Comments

Post a Comment

Popular posts from this blog

Where Mechanical At

Sprint 2 ended on Friday and we have our MVP! What that means for mechanical is that we have a box that can move nine different nodes, but control 3 sets independently, using three actuators. We made our prototype with some scrap .21" hardwood with the laser cutters. We converted the rotational motion of the sliders into linear motion with a pin and rack design, with the rack mounted on a slider with ball bearings to reduce the friction. The three gears were connected with a rod so that as one gear rotates the rest rotate as well. Our challenge coincided with retaining the gears with the bearings linearly, since the gear shaft would often tilt down (it was not a perfect press fit. We ended capping both sides of the shaft to retain it, but due to a misprint, the final design was switched, so the cap was not as effective as it should have been. The biggest issues with this design is that there was a lot of friction in the gears, and the system was too constrained, which made rotatin...
Read more

Mid Sprint 4!

The final sprint has come to pass and demo day is just around the corner! The mechanical team has worked hard to get the entire mechanical system up and running and the software/electrical team is currently working toe get the system flawlessly integrated with the web app and integrating the sound detection code into the system. We are really excited about the final demo day and we think we are in good shape for getting all of our ducks in a row in time. Stay tuned!
Read more
Image
Today Peter and I worked on reading and interpreting audio signals using pyaudio. We made use of functions from Allen Downey's ThinkDSP book to help us extract frequencies and magnitudes for .wav files. We were able to graph the distribution of the frequencies and their magnitudes and were able to see a difference when we recorded a sound with a low frequency and one with a higher frequency. Our next steps are going to involve figuring out how we can extract useful information (frequency with highest magnitude) as our code is running. We might also try to use threading so that we can be recording and interpreting sound data at the same time.  - Vicky
Read more