The charge seems to have been based on not the reporting itself, but a theory that he was conspiring with people committing crimes, like a lookout at a robbery. I wonder if it'd be on more solid ground if it were a website that just always updated with police movements, rather than only in connection with a specific event.
It's an interesting idea. Thanks to smartphones, the people have more cameras than their governments. The governments however have more targets.
Given a decent proportion of people using the app, you could probably map out the location of most of the active government vehicles in a city. With the time-series data, you could add some analytics and probably predict where the cars were going to be at specific times and days. You could even gameify it, so catching a police car at a donut shop would win the user a little badge.
Monetizing it might harm adoption by volunteers though, so that is important to consider. Gamification is a great idea though.
Inexpensive but ubiquitous and "always on" cameras, with a little bit of OpenCV smarts built in (or perhaps with a RaspberryPi-grade computer hanging off them) could probably manage a pretty good implementation of this.
If you could somehow roll this out as part of some innocuous looking project (off the cuff idea - approach a taxicab company, tell them you want to put camera's on each cab - so you feed each driver back information about nearby traffic density/speeds and approximate counts of people standing on sidewalks correlated with the number of other cabs around. Slip in ALPR without mentioning it, and a whitelist of marked and unmarked police car licence plates. Sell access to the real-time police location data to high-speed courier services, boy-racers, and drug cartels.)
The laws apply to us, not them.
The issue is that of technology being an enabler. Always being able to watch you (with some cost), and being able to watch everybody effortelsy and automatically, with instant retrieval is a game changer.