In this video I reflect on my initial experiences with OpenCV on the Raspberry Pi Model 5 under Bookworm.

I started earlier this month (Aug 24) with finding the best method of installing the latest version of OpenCV without needing to use a virtual environment on the Raspberry Pi. I used Method 3 in Reference 1 and successfully installed the program in a number of Raspberry Pi computers. Most Raspberry Pis had a 1TB SSD using the M.2 HAT. Besides learning how to use OpenCV I also wanted to learn how to use other libraries such as
* NumPy
* Matplotlib
* SciPy
* Scilit-Learn
* TensorFlow
* Keras
* DLib

I started with an unstructured approach and was able to get a couple of demos to work but had some problems with the others. Clearly I needed a more structured approach to my learning and adopted a book from Amazon (Reference 2: Mastering OpenCV with Python: ...). I purchased the Kindle version for $41.47. It is recent and covers the most recent OpenCV2 version you can install on the Raspberry Pi. There is also a website for the book and that saves you typing in all of the examples. In the video I will demo only a few programs from the book. You will have to purchase the book for details and the other examples.

While there are many errors in the code and text of the boo, I was able to correct theses and re-structure the images as sub-directories for each chapter to make running the programs easier. Despite these errors I recommend this book for simple examples only and doubt if I could construct the examples on my own. It saves a some time.

Other issues include that some of the images to not align with the example programs and some images need to be resized. Despite this I had lots of problems with the more complex examples. Programs for Chapters 8 to 12 did not work.

I am getting closer to more advanced examples from the book and will start Chapter 8 tomorrow with Machine Learning. So far I covered one chapter per day and have run all examples from Chapter 2 to 7 (about 70 programs). Had a quick look at Chapters 8 to 12 . Most examples in the book did not work for me. There were issues with python libraries like NumPy and some libraries were not previously mentioned or introduced.

In the next video I will reflect on my experiences with Chapters 8 to 12 and outline my future projects with OpenCV. After that I will explore video accelerators on the Raspberry Pi Model 5 for computer vision including the AI Hat for the Raspberry Pi and google Coral USB Accelerator

References:
1 Install OpenCV on Raspberry Pi 5 ( [ Ссылка ] )

2 Mastering OpenCV with Python: Use NumPy, Scikit, TensorFlow, and Matplotlib to learn Advanced algorithms for Machine Learning through a set of Practical Projects (English Edition) Paperback – 17 November 2023 ( [ Ссылка ] )

свернуть