This week I have continued to work on the seneye sensor. I have saved the information onto a file on PC and am currently working on getting it to work on the raspberry pi. Once this program works on a raspberry pi my project will be complete.

This past week I began testing with a video analysis and modeling tool. I used it to analyze a local high school student's lacrosse skills. I had him take a couple shots from about ten to fifteen yards away from the goal for me to analyze and give him constructive feedback. With this tool I was able to dissect the mechanics behind his shot and help evaluate his strengths and weaknesses. I was able to find areas for improvement so that he can develop his skills and improve t

Hello everyone, here is an update on the VARs station. When I last posted on the blog the VARs station checklist version 1 was in the process of going through validations. Since then the checklist has been updated based on the results from the v1 testing, and is now on version 3. This version of the checklist has finally passed the validations testing I guess the saying "third time is the charm" has some truth behind it. In all seriousness, I am very excited to see this v

How To Make A Shot List Making a shot list is crucial when making videos because it gives you a sense of direction. Without it, your videos will not be that good and you will just confuse your audience. It is not that hard to make a shot list all you need is a computer and create it in either Google Drive or Studio Binder. Below is an example of making a shot list using Google Drive and Filmsourcing. Also, there is a video that explains more in detail about a shot list This v

This week I started working on Garridae 190204. I started by sourcing parts from Bonefish class 223 and 582 and a drone autopilot testing board. Upon assembling the hardware, I had to configure the parameters. Because the thrusters only had a 30-second runtime in the air I mounted them to a testing board and submerged them in a bucket of water; this allowed me to tune the motor parameters while not burning up the motor bearings. At the end of the week, I took Garridae to the

This blog post is adopted from a lecture I watched by Chris Burns AI/ML Senior Solution Architect [1]. There are a few keys to building an image data set to train on the Sagemaker ResNet algorithm. First, one must collect Images. Many options are available here hopefully the client has a large collection of images for the objects in question. Image classification is dependent on clean abundant data source. One thousand images per class is a satisfactory minimum. Second, disca

This week I continued working on three projects, Gerridae 190204, Brushfire 190210, and Bonefish 190124. When working on the Gerridae 190204 project I started the week with a working differential thrust control system that had been pool tested the week before; on the autopilot side, I had a working compass but no GPS connection. I was able to rule out a hardware issue by swapping the Ublox module. I then was able to get the Ublox on Gerridae working by taking the parameters f

After research over nearly 2 weeks, including completion of Google's 'Machine Learning Crash Course' program (which explains and explores fundamental machine-learning concepts in an application-based setting), I will be starting work on training a neural net to recognize specific species of plants and animals native to the bay, and have it return information about the health of locally videographed ecosystems. Deployment of the LASR@ AACC program's Bonefish SAV for data colle

By Elisabeth Cawley This week I explored Amazon Sagemaker’s example Jupyter Notebooks. The example notebooks contain code that shows how to apply machine learning solutions by using Amazon SageMaker.[1] I tried a few of the image classification notebooks. They require Docker containers.[2] Docker Desktop Enterprise is a new commercial desktop offering that gives you everything you need for enterprise-ready container-based development, providing developers and IT the easiest,

This week I focused on building a remote controlled sailboat, and putting together two e-boards. The first eBoard that I created was the construction of the RC sailboat. Building it was fairly simple, except for rigging some of the lines. Creating this eBoard was quite challenging as I had to learn the software. The second eBoard instructs how to operate the RC sailboat with the given controller. Both will be in validation towards the end of the coming week. I also assisted M