Jim, as valid as your points are - if you don't search in the right area, you won't find what you are looking for. .. ...
Yes.
In both cases, the search system you describe and the proposed spectral imaging search tool, there are two major shortcomings to account for. First, the processing capacity of the system, and second, the quality of the mathematical model. These interact because excessive complexity of the model eats into your processing capacity. The trick is to create a mathematical model that harvests the maximum effectiveness without excessive amounts of code. The commercial reality is that with budgets and timescales to meet then if your first shot at it is not quite right the next try is going to be a half-baked add-on that makes the right lamps come on but doesn't do the job properly. By the time you persuade the customer to pay for Mark 2, the guy that wrote the code has moved on and taken his notes with him. Sound familiar? I can tell you it sounds really very very familiar to me in my engineering world!