In social science experiments, it's not unusual to have undergrads or grad students classify the raw data for certain characteristics (a process generally referred to as coding). If we produce (the crowd) a large amount of data, do you really think they'll reject classification information? If someone walked into their office and said "Look at this picture I took of a guy putting something in a trashcan", I suspect that person would get pretty immediate attention.
These agencies often appear to be slow because they've got political bureaucracies as well as defined processes (chain of custody in evidence is pretty important), but it's a mistake to conflate their slower speed with a lack of intelligence.