In order to give the end user the ability to tune the software specifically to their own requirements we recommend the amazingly useful
open source application for cellular imaging - OpenCFU developed by Quentin Geissmann.
The application is free and open source, and easy to configure. Its functionality sits well with our machines as most of our own applications are also written in C/C++ and we use GTK+ software for designing GUI's and custom interfaces.
This application, combined with a low cost webcam (such as our 5MP raspberry pi camera attached to our Gilson 215 liquid handler) can accurately and quickly count colonies and determine which have the highest rate of growth, and we can export the results directly to SQL or other scientific data repositories such as sciNote and automatically identify the fastest rates of growth as its happens.
There is a research article describing the implementation and the algortihm is available at journals.plos.org.