Detecting Fights In Prison Bathrooms: How Would You Do It?

Actual application question we received: a medium security prison is having incidents in their bathrooms, mostly fights in the shower area. They cannot put cameras in the room, more or less (it's possible but very touchy and requires a lot of paperwork and regulation).

They can put a camera outside the room facing the door, which looks similar to below, which would get them good video of who is entering/exiting. They can shoot through the doorway and window as well, since those areas are technically outside of the shower/toilet areas. But they would be so obscured I don't think video would be the most useful thing here.

There is also discussion of doing audio analytics/alarms, like the Louroe Aggression Detector. I thought it may be worth putting a PIR in the room as well to notify when it was actually occupied, so the VMS (Genetec) could switch views and guards could be more aware.

Thoughts on this approach/approaches? Any other ideas?


This can potentially be a good application for a thermal camera, possibly with crowd-detection analytics. The thermal camera will prevent the capture of identifying details like faces, or even tattoos. This is a case of lower resolution being better, so one of FLIR's Lepton cameras could be ideal here.

They could also use an analytics camera with an alarm out to signal when a fight may be occurring (again, via crowd detection event), and just feed an alarm I/O signal into Genetec to alert the guards when they should go into the shower room to investigate in person.

LIDAR

Louroe guy here.

The aggression detector mentioned will send an alarm to guards or security staff when there is yelling or fighting. Regarding privacy, the detector does not listen for words or phrases and it is not necessary to record or listen to an audio stream. In practice a guard will receive an alarm that an aggression event is taking place in the shower area without needing to monitor the shower area 24/7.

Cameron, is this theoretical or have you done a project like this? If the latter, can you give more details about what / how it worked?

Would the background noise of the shower spray cause problems for analyzing the audio?

The aggression detector has been used in prison & interview room applications, so there is real world experience beyond the lab. Typically the detector runs on the Axis ACAP platform as shown in the IPVM tests. In this case where cameras won't be in the shower room, you could use the Axis F41 module and not connect a lens to it. From there several options are available on VMS integration depending on what the customer wants and the specific VMS being used. Most customers have a notification pop up at a guard station, with audio recorded pre and post alarm. The extent of the VMS integration depends on which VMS is used and with the majors there are very few issues. Other customers have used relay outputs from the camera or F41 to turn on floodlights or lock doors.

Brian, the shower spray likely wouldn't be an issue unless it's at a very high DB level and close to the microphone. The detector threshold adjusts automatically to changes in background noise and I'm assuming most fights would be louder than the noise a shower spray makes.

The thermal camera is not a bad idea. but then your resolution is always low and it's an expensive option. Since we are a Geutebruck dealer, we'd suggest using a camera and Geutebruck's Motion on Privacy analytic which either blurs or pixelates people moving in the picture. It would protect privacy while maintaining high resolution of the rest of the area.

FLIR thermal cameras are a couple hundred bucks. The problem is the fact that they aren't even a little vandal resistant.

Point the camera at the soap dish?

We should put Ethan is a prison bathroom for a day and see what he comes up with ;)

The inmates will need to be chipped with technology that broadcasts heart rate, skeletal motion/movement and even audio transcripts of every word spoken. Perhaps even the actual thought patterns of behavior and emotion can be transcoded to the matrix.

Freedom and rights need to be reconfigured to transcript the human life, we are still in the stone age of technology.

How about FLIR cameras would show images without giving private details

Micheal, the Flir and thermal were already mentioned.

People counting analytic. 4 go in, 3 come out.

We were able to put cameras in the bathroom but they could not capture the actual toilet or the shower, so they are installed so that everything can be seen except those two areas of which are not a large enough area to have a fight. However, the shower area posed a different problem in that it would not be able to capture if two persons went in the stall for "other" reasons around the PREA. One area has a corner mount vandal no grab mount by video alarm and we rigged in an axis p1214 angled and positioned directly above the toilet and pointing out. You can't see the toilet but can see when they walk up to it. I suppose it all comes down to each ones layout.

How Mexico’s Most-Wanted Drug Lord Escaped From Prison (Again)

I believe El Chapo escaped through a hole under his shower, where the camera could not look due to privacy rights issues. That's why either a thermal camera like Brian suggested, or in my mind better yet, a full resolution camera that pixelates a person's image in the picture so you can see what is going on without the privacy details, and make it based on the operator's login rights whether they see the pixelated or non-pixelated video; ie. a prisoner does get attacked and you want the full res video for prosecution.