There are a few other considerations. Actually building a robot that can do even a poor but detectable job of finding bomblets (or anything else) is an extraordinarily difficult task. If you achieve it, you will become wealthy and powerful!
Alternatively, you can have the kids do more manageable subsets of the problem, for example write computer programs that examine photographs and draw circles around objects they think are bombs. The program can look for things that are of a certain size and color, maybe with hints regarding the lens used on the camera and the distance to the object in focus so that the program has an idea of the scale of the objects it is examining. If you have the resources, you can have them build a Lego robot with a camera on it, optionally hooked up to an off-board computer, that can do a similar analysis of what the robot sees through its camera.
The actual robot 1) doesn't need to have all the compute power on-board and can offload it to a nearby computer over a wire or WiFi network, and 2) does not even need to physically exist, students can write not-so-complicated programs to brainstorm and test interesting strategies for object detection.
How old are the students? How much computer programming can be realistically expected from them?