PTZ Best Tracking Tips

[TVille]

I got my @EMPIRETECANDY 49225 yesterday. Damn its big and heavy! I have it on a temporary mount - a 4' 2x4 in a bucket of gravel, as @looney2ns recommends, and it still shakes some on max zoom! Anyways, it works great - zoom is incredible. If you don't have one, get it now!

I read @wittaj's tips on how to set up spotter cams for a PTZ. This is working well for my test location, I currently have a pair of cams working as spotter cams.

Any other recommendations? Like adjusting some of the more obscure things such as track time, tracking target size, etc. One issue I have right now is trees in the way. I have several larger trees or branches in the way which stops the tracking as the subject goes behind the trees. I also am a considerable distance from the subjects, they range from 80' to 375' away, which is where the zoom comes in handy!

 

[Smilingreen]

I got my @EMPIRETECANDY 49225 yesterday. Damn its big and heavy! I have it on a temporary mount - a 4' 2x4 in a bucket of gravel, as @looney2ns recommends, and it still shakes some on max zoom! Anyways, it works great - zoom is incredible. If you don't have one, get it now!

I read @wittaj's tips on how to set up spotter cams for a PTZ. This is working well for my test location, I currently have a pair of cams working as spotter cams.

Any other recommendations? Like adjusting some of the more obscure things such as track time, tracking target size, etc. One issue I have right now is trees in the way. I have several larger trees or branches in the way which stops the tracking as the subject goes behind the trees. I also am a considerable distance from the subjects, they range from 80' to 375' away, which is where the zoom comes in handy!

Pictures! You can't buy a new toy, brag about it and not show pictures to assist in rubbing it in!

 

[wittaj]

General tips:

As with every camera, dialing in the parameters for brightness, contrast, shutter is important, but even more so with an auto track PTZ or it will lose track and stop on mailboxes and other stationary objects.

Target size is a balance of getting the tight shot you want yet not be so large that it loses track. I'd recommend starting at 35 and then adjust from there so that it tracks around obstacles.

The spotter cams can also work well to redirect the cam if it gets blinded by the tree so that it can move PTZ to other side of the tree.

 

[looney2ns]

I got my @EMPIRETECANDY 49225 yesterday. Damn its big and heavy! I have it on a temporary mount - a 4' 2x4 in a bucket of gravel, as @looney2ns recommends, and it still shakes some on max zoom! Anyways, it works great - zoom is incredible. If you don't have one, get it now!

I read @wittaj's tips on how to set up spotter cams for a PTZ. This is working well for my test location, I currently have a pair of cams working as spotter cams.

Any other recommendations? Like adjusting some of the more obscure things such as track time, tracking target size, etc. One issue I have right now is trees in the way. I have several larger trees or branches in the way which stops the tracking as the subject goes behind the trees. I also am a considerable distance from the subjects, they range from 80' to 375' away, which is where the zoom comes in handy!

It has to be an extremely stable mounting to prevent shake, you may not eliminate it totally at max zoom, especially on windy days.

 

[EMPIRETECANDY]

That is the smallest and light 25X zoom ptz, the huge one 28.66lb

 

[TVille]

That is the smallest and light 25X zoom ptz, the huge one 28.66lb

Yeah, this is only like 10 lbs. But, compared to a 5442 turret, it is huge!!

 

[sebastiantombs]

Compared to a 55 gallon drum it's small. Ya gotta look at it from the right perspective

 

[TVille]

Compared to a 55 gallon drum it's small. Ya gotta look at it from the right perspective

True! And the wife has not complained about the size, even sitting 4' above the front porch, so it's not that big!

But, it does one heck of a job looking 1,000' down the street, even at night!

 

[wittaj]

Yep, my 49225 replaced my first PTZ, which was a non-tracking PTZ. I thought that thing was huge and the 49225 was double LOL.

But the 49225 is such an incredible unit. Great range and with spotter cams, an amazing addition to your coverage.

 

[bp2008]

I just played with the auto tracking on my new PTZ, and so far it is impressive ... in how much it sucks.

I have 3 main complaints.

1. Awful target acquisition success rate.

It works maybe 10% of the time, which means a 90% rate of failure to detect me or a vehicle crossing any of the triplines or intrusion zones I configured. At least it is this bad at night.

First I just tried one big intrusion zone filling the entire field of view.

Configured like this:




It was awful. I walked through the scene from one corner to the other and it didn't even twitch. Walking back into the scene from the left, it noticed me and tracked me.

So I replaced the intrusion rule with a tripwire rule where I drew a box around the outer ring, and a second box inset a little in case it needed to have me fully in the picture before I crossed the line. It was awful. No trigger at all.

I redrew the tripwire rule as a big spider web of lines systematically covering the entire view. You couldn't walk two feet without crossing a line. So I went outside, crossed the FOV a few times. Walked to the center and back out. Not tracked even once. Complete garbage.

Finally I deleted that rule and made 5 new very simple tripwire rules and 3 very simple intrusion rules that cover basically everything. Once again, very disappointing. As I walked down my steps, the camera decided to tilt straight up away from me, then back down again a second later. I got it to follow me properly two times, but at this point it is becoming obvious that the "Deep IVS" detection is just trash. At least it is at night. Maybe there are some tricks to make it more reliable. I don't know. I feel like I tried everything that is reasonable already. Several times, I saw that camera pan and tilt over to the road, pause for literally one second, and return to preset 1. Two of the times there wasn't even anything there. Once, a neighbor was pulling out of their garage. Either way the camera didn't properly track anything.

2. Dumb Track Time setting

You can only give it a fixed number of seconds to track a target, and once that time expires it will return to its preset position whether the target is still in the middle of the view or if they escaped the view ages ago. No intelligence whatsoever.

3. Target Tracking Size

As for the target tracking size, I found that setting it to the minimum value of 10 made it not zoom at all when I walked around in my backyard. A value of 50 was pretty decent, zooming to capture my whole body with a little bit of margin above and below. A value of 75 was a bit too much, as my head poked out of the field of view often. This camera supposedly is using deep learning human detection, and I guess it doesn't even know the head is at the top of the body. It was evidently trying to center itself on the center of my body.

 

[bradner]

True! And the wife has not complained about the size

a size queen........nothing wrong with that

 

[bradner]

I just played with the auto tracking on my new PTZ, and so far it is impressive ... in how much it sucks.

I have 3 main complaints.

1. Awful target acquisition success rate.

It works maybe 10% of the time, which means a 90% rate of failure to detect me or a vehicle crossing any of the triplines or intrusion zones I configured. At least it is this bad at night.

My two PTZ's autotracked much better - not great overall but much better - by not using an IVS functions.

 

[TVille]

@bp2008 have you tried setting the trip wire or box inside the frame? By say 10% of the FoV. I have 2 trees blocking my temp setup, and only 180 degrees view, and it works fairly well at catching motion. Tracking does kind of suck given the two large obstructions. I also have two spotter cams.

Sent from my Pixel 4a using Tapatalk

 

[sebastiantombs]

I'd move the edges of the intrusion box further into the frame. I think cameras, generally, like to see something in the FOV before it enters the detection zone.

 

[wittaj]

@bp2008 - as you are seeing, these are not plug and play. You have to dial them in. But once you do, they are incredible. Mine hits the mark over 95% of the time.

Are you running default/auto settings?

It is a matter of getting the brightness/contrast and target ratio settings correct.

Looking at your image above, there is a lot of gray (yeah I know it is B/W lol) so it all blends in the eyes of the PTZ. Adjusting the parameters can help give it a little bit of contrast so that it has an easier time tracking.

I always knew that you shouldn't chase a bright picture - it looks nice and people migrate towards a brighter TV for example, but upon closer examination, most images need to be toned down in order to get all the details. You will be surprised how much changing a parameter like gamma could impact tracking. For example, if you have a pesky tree or something in the middle of the view during an autotrack, just by changing some image parameters you can get autotrack to pass it. Making the image a little darker at night actually helped with tracking someone across the street, which was opposite of what I thought you would think to do. So add some contrast to your image above and see if it improves.

I have a yard lamp post that more times than not autotrack would get stuck on it as someone was walking and the autotrack would only go so far. Because my image has soo much contrast (bright white concrete a third, blacktop road a third, grass a third), knocking down the gamma made the lamp post not be so "trackable" lol, and along with that I turned of PFA and that gave it just enough time to retrack the person walking past the lamp post. The camera may still autotrack the lamp post when a small kid goes by, but an adult it was autotracking past the lamp post.

A big issue I had when I first got mine was I couldn't get the autotrack to zoom in closer than 7x. Just by changing your initial FOV and adjusting IVS size to give it some headroom made a big difference. Not using a minimum target size made a big difference - keep it at 0,0 (which it looks like you are doing) and let AI do its thing. This thing will now zoom in 25x and have a nice tight shot of someone hundreds of feet away.

After you do the above, it will start to work better. To further address your big 3 complaints:

1. Awful target acquisition success rate.

Ideally for an intrusion box or tripwires, you should have the initial field of view be such that the camera doesn't have to initially pan too much up/down or left/right to get the object in the center of the screen to start tracking. The closer the object is to the center of the image, the better the chance that it will track correctly.

The reason it starts looking upward or left or right is usually because the intrusion box is too big so the camera identifies the object before it is in the center of the field of view and then sometimes something else matches the "algorithm signature" of the initial object and then starts trying to track something that isn't there. Adjusting the field of view and the locations of the IVS rules to be closer to the center can fix that.

Alternatively, you can set the PTZ to limit the up range so that it won't do that, but I would prefer dialing it in first and only do this if it still happens.

2. Dumb Track Time setting

This one is simple - simply make it 300 seconds. As soon as the object is out of the field of view (probably within 300 seconds) it swings back to home.

This PTZ does need to be set in Explorer or you can only set this time to 15 seconds. It will revert back to 15 in any other browser.

3. Target Tracking Size

Make sure you don't give a min object size - keep it at 0,0. Looking at your initial field of view, you need to drop the target tracking size lower. Try 35 and test it and move up and down from there. And then give it some time. After about a week, mine started tracking more the head than the body.

 

[bp2008]

Thanks for the tips everyone, especially @wittaj with the information dump.

I did leave the image settings all on auto, and I was suspicious that higher contrast may be required as it mostly seemed to notice me near the center of the frame where the IR is strongest. I'm going to see how auto image settings works during the day, and then consider whether I want to try improving performance at night via a night profile. For the most part there will be no obstructions like trees or lamp posts getting between the camera and people, so it sounds like adjusting the image should be fairly straightforward.

Even if I do get the tracking working well, I may end up just leaving it off anyway because the chance of it being useful in my backyard is next to zero, while the chance of it malfunctioning and causing recording spam is much higher (so far. heh).

 

[bradner]

@wittaj, after reading your post i may revised using IVS for my PTZ cams. Thanks for the info - gamma I've found a bit confusing on how to adjust it.

 

[wittaj]

After you dial it in, I suspect it will be rock solid in your field of view and you can let autotrack do its thing without all the spam!

 

[Gargoile]

@wittaj Good points and thank you for your posts. I too just received this same camera. I have this stilling on my computer upside down playing with all the settings. This is my first PTZ so I guess I am spoiled.


.

 

[TVille]

This PTZ does need to be set in Explorer or you can only set this time to 15 seconds. It will revert back to 15 in any other browser.

As usual, your comments are dead on, thanks!

I have been using Brave (a Chrome browser) under Linux Mint for all of my Dahua cameras, and it works with them all. Until now. I checked, and you are mostly right. I can change the tracking time under Linux, but it shows 15 seconds. Go into IE, and it shows 30 seconds, what I had changed one of my presets to, under Linux. Back to Linux, refresh the page with the refresh button, still 15. So, it almost works, your just blind on that setting.