The elephant detection system will not be finished in time I'm afraid! The biggest problem towards the end was not being able to access elephants in order to obtain training images! The best training images would have come from the raspberry pi camera,...
For the last part of the Computer Lab Box I wanted to have the possibility to make a net boot provided by the main computer. This was a very easy task. I found already before a good tutorial on https://bootlin.com/doc/training/linux-kernel/. For now,...
After adding about 1,000 'labels' as described in the previous log, rather surprisingly, the detection now works very well in bright sunlight with strong shadows: It can be seen that the blue detection boxes are tighter on the plants than before. Here's...
- Continued training SceneTextRecognition AI to read, which didn't go exactly in our favor, as AI training can take months to complete- Tested our newly made box for its efficiency, made sure that it fitted correctly before our upcoming presentation...
So, this week I decided to test the TinyDecisionTreeClassifier with my fan to see if it can recognize the 'broken' fan movement. Nothing fancy, one minute of training data for each state(ok/failure), then few seconds of training process, then it...
I finished putting all the pieces together and wrote a little code to tie it all together. It waits 5s after a motion event to let the cat get nice and settled, triggers a 1s burst, waits for 10s to let it clear the field of view and let the PIR settle,...
The slugs are now house trained. 10605 images of slugs have created a fairly good slug detector although there could be improvements by adding blurred images to get the training to focus more on the overall shape of the animals, including their tentacles,...
I finally received the new speed controllers. They're now wired up to the Pi through a quick resistor bridge I made to do the DAC conversion for the throttle.From here on out the project will be mostly focused on coding the web server and training the...
To control the proposed motorized exoskeleton glove, the wearable muscle-machine interface was used. It combines Electromyography (EMG) and Forcemyography (FMG) sensors to decode the user’s intentions and discriminates between different grasp types....
Many more enhancements, important corrections, oversights, refactorings... and realizing that another things I needed is a long training and seeing the reward plot specially.By the moment it's looking good and seems that learn correctly instead of unlearning...
So squeezing the training data to match animorphic 16:9 video didn't give any hits. When the test video was cropped to 1:1 again, hits bounced back to the same as if the training data was never squeezed. It somehow knew the test video was...
I was able to get live predictions with only 5 min of training data with the sensor system.The accuracy is lower but I just wanted to make sure that I had done everything correct.
Here's a list of things that need to be done for the project: Add build instructions.Develop the software and put it on Github.Make the device's fabrication easy: make it out of cardboard.Choose the sensors (piezo).Design the sensors interface circuit,...
I have decided to first develop a simple DC supply voltage controller that will be the basis for two configurations described in the introduction (controller acting as a voltage source for the self sustained DC "grid"; controller connected to the grid-tied...
-shhhhhh- ...... I'm saturated with school work in my grad program, so I'm not going to be able to turn much attention to rewriting the instructions and such for a bit, and the formal release with all the pageantry and publicity and awesome video...