See-TreePiO

I researched more about the tree datasets, and I found one that has plenty of accurate information: PlantNet API.

I researched the different datasets I can use for the project, as well as the amount of tree data each have in them. I still need to do some more research, but as of now, creating a dataset is the best option if I want to get a wide, accurate range of data. I haven’t found any datasets that match the amount of data I can collect by myself in my neighborhood.

I tested the camera and battery pack that came. They both worked really well. Now, I can start inputting the ML algrothm.

I fiinshed the Machine Learning book.

I read Chapters 14 and 15 of the Machine Learning book.

The parts will come tomorrow.

I read Chapter 13 of the Machine Learning book.

I planned the way I am going to connect the ArduCam to the Raspberry Pi with the thermal sensor. I realized I might have to solder some wires together so that more than one wire can fit on one pin on the Pi.

I planned the way the wires are going to connect to the Pi and how I am going to integrate ML into the Pi using the camera and thermal sensor.

I researched more about using the ArduCam with the Raspberry Pi 5. I found troubleshooting techniques and what I need to do after I get the camera (the setup).

I read Chapters 11 and 12 of the Machine Learning book.

I watched a video about using the ArduCam with the Raspberry Pi 5.

I read Chapters 9 and 10 of the Machine Learning book.

I read Chapter 8 of the Machine Learning book.

I will stop watching the AI on the Edge videos because they are not accessible on YouTube after Lesson 31.

The parts will come day after tomorrow, so I plan to finish reading this book by then.

I watched Lesson 31 of the AI on the Edge videos.

I watched Lesson 30 of the AI on the Edge videos.

I watched Lesson 29 of the AI on the Edge videos.

I watched Lesson 28 of the AI on the Edge videos.

I watched Lesson 27 of the AI on the Edge videos.

I watched Lesson 26 of the AI on the Edge videos.

I watched Lesson 25 of the AI on the Edge videos.

I watched Lesson 24 of the AI on the Edge videos.

I read chapters 6 and 7 of the Machine Learning book.
I’m just reading and watching the videos until the parts arrive in three or four days. Then, I will work on the technical aspect of the prototype. I have some free time now, so I thought I would learn about Machine Learning and using the Raspberry Pi in the time I have instead of wasting it.

I watched Lesson 23 of the AI on the Edge videos.

I watched Lesson 22 of the AI on the Edge videos.

I watched Lesson 21 of the AI on the Edge videos.

I watched Lesson 20 of the AI on the Edge videos.

I watched Lesson 19 of the AI on the Edge videos.

I watched Lesson 18 of the AI on the Edge videos.

I watched Lesson 17 of the AI on the Edge videos.

I watched Lesson 16 of the AI on the Edge videos.

I watched Lesson 15 of the AI on the Edge videos.

I watched Lesson 14 of the AI on the Edge videos.

I watched Lesson 13 of the AI on the Edge videos.

I watched Lesson 12 of the AI on the Edge videos.

I read chapters 5 and 6 of the Machine Learning book.

I watched Lesson 11 of the AI on the Edge videos.

I watched Lesson 10 of the AI on the Edge videos.

I watched Lesson 9 of the AI on the Edge videos.

I watched Lesson 8 of the AI on the Edge videos.

I finished reading Chapter 4 of the Machine Learning book. I am also taking more notes on the book now.

I sent an email to Ms. Mint (the one I met previosuly) asking if we could meet again for some feedback on my project idea. I also converted my paper plan onto a document and calendar so I can keep track of any deadlines I might have for the project.

I watched Lesson 7 of the AI on the Edge videos.

I watched Lesson 6 of the AI on the Edge videos.

I watched Lesson 5 of AI on the Edge. In the previous DevLog about the AI on the Edge videos, I actually watched Lesson 4, not Lesson 3.

I made a plan on how to spend my time during this summer, as a lot of time will be dedicated to the ML aspect of my project.

I watched Lesson 3 of the AI on the Edge videos.

I finished the 3rd lesson of the AI on the Edge.

I watched the Lesson 2 of the AI on the Edge Raspberry Pi videos.

I watched a video about using the Raspberry Pi 5 with ML. Though the Machine Learning book is helpful (and I will keep reading it), I felt that it was not enough to help me with this specific project.

I researched exactly what camera and battery pack to buy (to power the Raspberry Pi). I added them to the cart and will buy it as soon as possible to further develop the prototype. I can’t wait to see how it works with the thermal sensor!

I just researched about how to combine a thermal sensor with a camera. I found a helpful video and took notes on that video. I also found the specific type of camera I can use, which is a standard Raspberry Pi Camera Module, as well as the connector.

Instead of physically testing the prototypes out (because my dog accidentally stepped on the ultrasonic sensor prototype), I just researched a bit more about the two sensors. I researched the benefits and downsides of each sensor, as well as how accurae they are at specifcally detecting dead trees. I found that people have used ultrasonic sensors before to detect dead trees, but no one has used thermal sensors in the way that I am before to do so. Therefore, I will stick with using a thermal sensor for this prototype.

I will have to do more research about the actual pairing between a thermal sensor, a camera, and ML, but the book will help me a lot in the coding of the ML portion.

I compared the 2 prototypes to see which sensor is more accurate (with research, not actual testing). I found that the thermal sensor is more accurate, especially when paired with a camera and an ML algorithm. However, the pairing will be new to me because of the Raspberry Pi. I am more used to working with Aruinos, but through research and practice, I will be able to implement ML. Also, the book I am reading is going to help increase my knowledge on ML.

In a future DevLog, I might just test this research just to be sure the thermal sensor is the right choice.

I created the new prototype using the ultrasonic sensor. I coded it, and it successfully functions properly. However, when I tested it with a tree, it didn’t detect it correctly (it said it was dead, not healthy). When I tested it a few more times, it showed accurate readings.

I still need to compare the two prototypes together.

I started taking notes on how to work with a thermal sensor, a camera, and ML. I found that it might be hard to integrate all 3 of these things, so I started looking in a different path.

I found that I could use ultrasonic distance sensors to detect the unique ultrasonic waves emitted from trees. Therefore, I will start creating a prototype using the ultrasonic sensor and compare the 2 prototypes together.

I copied the code from Command Prompt to GitHub so I have a way to store it somewhere safe. I also double-checkked that the code was correct, and fixed any errors.

I interviewed a tree expert, Ms. Mint, who works at a local farm. She brought up some interesting points, and helped me narrow down what sensors I want to use for my project. She said that using a thermal sensor would be a good idea becuase of its accuracy and easy use.

However, when I interviewed her, I had to record the video on my phone instead of on this Hackatime. Therefore, I screen recorded the video and that is the video shown here on this Devlog.

I was also with a few other people interested in Ms. Mint’s work, so it wasn’t a one-on-one session, more of a broader overview with a lot of Q-and-A.

Overall, this experience was really nice because I got to meet a curator of a nature center, feel the nature all around me, see the difference between dead and healthy trees, and get feedback on my idea.

I wired the thermal sensor, the LCD, the Raspberry Pi, and the LCD backpack. At first, the wiring was not working, so I switched to planning the wiring on paper. Once I finished that, I used a soldering iron to combine two wires to fit in one slot. This is the completed wiring.

I continued reading the Machine Learning book and got to Chapter 4. The book is really interesting and I already feel like I have some knowledge (not a lot, but some) about ML.

I continued reading the Machine Learning book and got to Chapter 4. The book is really interesting and I already feel like I have some knowledge (not a lot, but some) about ML.

I have started reading a book about Machine Learning, its history, and how to train/use it to the best of my ability. I have been taking some notes and answering some questions presented in the book. I have gotten through the first 2 chapters so far and plan to finish the book in the next few weeks. This book will really help me use ML in my project, as well as use it in other projects that I may decide to pursue.

I created the code for the basic part of the prototype. The code makes the prototype able to detect the thermal signature of a Redwood Tree and determine whether it is dead or healthy. In the future, I wil re-make the code to incorporate a camera and a Machine Learning algorithm that can be used for more accurate and broader readings, not just for Redwood trees.

Researched about the different techniques and sensors I could use for this project. I found that using a thermal sensor with Machine Learning algorithms and a camera would be the best for my project.

Testing and editing the code and the wiring of the prototype