You need a library to recognize the machine numbers.
There is a choice between two libraries Keras and OpenCV
Which is better to choose? Or is there an even better solution?
Programming language Python3
This two-part article by Trung Tran about Real Time Object Recognition demonstrates differences between Keras and OpenCV from a coding point of view is a reasonable treatment. It also covers implementing pre-trained VGG16 models.
The application in this Campus Hippo lesson uses both Keras and OpenCV to simulate a self-driving car.
This tutorial by Adrian Rosebrockis about installing Keras for deep learning contains some good installation information to get both Keras and OpenCV running in the same virtual environment, which may be useful, especially for initial experimentation.
Being a user of Python3, I never limit my options to Python3 libraries. Python is used as a scripting language in place of Bash scripts in operating systems today because it can call Javac, C or C++ executables, or any other library or executable and pass values, arrays, or matrices back and forth with ease and simplicity.
Regarding choice, you may wish to define what better means in your context. If you wished to quantify or at least qualitatively describe acceptance or comparison criteria, there are a few typical dimensions to consider in your selection process.
Tangential Comment: For clarity, you may wish to specify more information about the machine numbers in the question. We can only guess whether you mean some numbers rasterized using some font and kerning rules.