We developed software to solve the problem of counting growing plants

I'm a bit perplexed by this. On its surface, it's a case study being used for Evrone (a Russian software development firm), presumably to convey their expertise. However, there's essentially no meaningful detail supplied. The problem itself is fairly well-stated, but the useful information stops about there.

The page claims that "there are several ready-made solutions in the market that could be easily integrated with the system developed by Evrone" but does not describe how the product was tested, what their results were, under what conditions it was tested, what hardware was used, nor provide a comparison of those results to using the traditional methods--all things that I know I'd want to consider before hiring a contractor for development work or further considering this particular technology into an application.

And then their "used technologies" section simply lists "Rust, Python, Artificial intelligence, Machine Learning", with no other information. While I don't believe the use of Rust in this particular problem space is common, there's not nearly enough information to form any sort of conclusion, and saying that the stack contains Python and AI/machine learning is again exceedingly generic.

I’m skeptical at the method of using video images to create what is essentially an orthomosaic. Either this technology is BS or this article was written by someone who is decidedly non-technical.

Great job!