Here we go… I put myself to the task of identifying/collecting places in my city and proximities that feature a certain kind of handwritten/handmade signs. The best way I find is to work with Google Street View and navigate through the streets and save/bookmark the place in Google Maps every time I find the kind of sign I want. But this is too time-consuming and I’m trying to find a way to make this more efficient.

I’ve downloaded a software called “Street View Download” where I can set an area in the map and it downloads thousands of Google Street View Panoramas from the area. This worked amazingly.

So I thought I could run the images through some OCR software and then I could search for keywords that I found in the signs I wanted. I’m using an OCR program called “Search for text in pictures” that in theory can do this exactly the way I need, but the results are very underwhelming. It barely recognizes the text from handmade signs. I’ve tried some online tools that can recognize text much better, but unfortunately, those are not made for huge batches of thousands of images.

I wonder if anyone can help me with a way to make this work better. I’ve heard of models you can train by providing the correct text and things like this, but I’m inexperienced with programming, coding, or complex tech. Is this too difficult to implement? Am I wasting my time and should just go back to the good ‘ol Google Street View navigation?