Hi @Luuk_Huizinga!
This is an incredibly exciting project and a great use case for Roboflow! Five images with 300 annotations is a great start, but we suggest you annotate 50 images, do some augmentations and then train a model. It’ll be pretty good.
If you want to improve from there, we suggest a couple hundred. Worst case, we do not expect you to need more than ~1,000 for a viable project. Roboflow has a nice article on how many images companies tend to use - Computer Vision Trends Report 2025 - Key Benchmarks (Keep in mind that the few companies using thousands of images likely have some major manufacturing processes that require high accuracy for critical operations. You might not need to be that robust.)
My suggestion is to implement an object detection model such as yolov8 , which you can train tune on your annotated symbols. You may have come across this in your research, but to reiterate, ensure that your training set includes edge cases such as tightly grouped labels to help improve model accuracy. Also, fine-tuning the anchor box sizes and confidence thresholds can further help the model reduce false positives.
Here is a link to our docs on training models: Train a Model
For example, here is a high level overview of a workflow that you could implement. Using the Roboflow Dataset Upload allows you to quickly review predictions, accept or reject them, and add to your training set with new labeled examples.
Please keep us updated on your progress and reach out with any further questions!