Documentation and loose ends
Goals from This Past Week
Team goals: Assemble a physical product and improve performance through testing.
Personal goals: Begin drafting final report
Accomplishments
I started off this week by working on drafting our final report. Since I was waiting for Tyler to finish printing the enclosure before working on more hardware, working on the final report was a good way to remain productive.
In researching other SET classifiers, I did not find any other systems that aimed to be used in embedded systems. Most existing algorithms extensively use OpenCV to implement their algorithms with the assumption that the systems running the code have access to a full Python distribution. Use of extensive OpenCV makes these existing classifiers not as efficient as they could be, especially since we found that some of the classification tasks ended up having very simple solutions. I think our project is a very good exercise in trying to achieve the same accuracy as these other systems while using only a fraction of the computational power.
It was also interesting to do more research on the theory behind all our algorithms. I knew generally what was going on with these algorithms thanks to prior computer vision and machine learning courses, but having to do the research again was helpful for solidifying the theory.
Hardware
In developing our physical system, we realized that a power switch would be very helpful for the user. Repeatedly plugging and unplugging the battery from the JST connector in the OpenMV board would be both tedious for the user and wear out the connector over time. So, I created a simple switch that sits between the battery and main board so it is easy to turn the system on and off.
At this stage it still needed to be crimped to the JST connector, but Tyler is the one who has the proper crimper. Here is what it all looks like together after crimping.
Goals for Next Week
Team goals: Prepare and practice final presentation
Personal goals: Finish a draft of the final report
Kommentarer