01304 827609 info@use-ip.co.uk Find us

Motion detection or intrusion?

alexdel

Member
Messages
18
Points
1
Hi,
I would like to ask a silly question that I can not answer to my self.
What is the actual difference between intrusion detection on smart events vs simple motion detection?
In both cases there is the human and vehicle verification, the draw area option as also the linkage method.
Thank you
 
Thanks @alexdel

motion detection and intrusion detection do trigger in a similar way (e.g. object enters a set area to trigger the alarm), the difference is the control over how the detection activates.

Motion detection is a basic blunt tool, any moving object that enters the detection area will trigger the event. You can adjust the sensitivity (the higher the sensitivity the more sensitive the detection is) and set human/vehicle, but even with these set, the event will trigger for most human/vehicle motion.

Intrusion detection has a little more fine adjustment to it, the initial set-up is the same in that you draw detection areas but that is the only similarity. After setting detection areas you can then set minimum & maximum detection sizes so that small animals or very large vehicles don't trigger the event, you also can adjust sensitivity and threshold which are explained below:

Sensitivity
Sensitivity stands for the percentage of the body part of an acceptable target that enters the
Detection region. Sensitivity = 100 - S1/ST × 100. S1 stands for the target body part that goes
across the Detection region. ST stands for the complete target body. The higher the value of
Sensitivity is, the more easily the alarm can be triggered.

Threshold
Threshold stands for the threshold for the time of the object loitering in the region. If the
time that one object stays exceeds the threshold, the alarm is triggered. The larger the value
of the threshold is, the longer the alarm triggering time is.

So for example, if you had an intrusion detection area set on your driveway and it stretched slightly onto the pavement at the end of your drive, you could set the threshold longer so members of the public just walking past quickly don't trigger the event.
 
Thanks @alexdel

motion detection and intrusion detection do trigger in a similar way (e.g. object enters a set area to trigger the alarm), the difference is the control over how the detection activates.

Motion detection is a basic blunt tool, any moving object that enters the detection area will trigger the event. You can adjust the sensitivity (the higher the sensitivity the more sensitive the detection is) and set human/vehicle, but even with these set, the event will trigger for most human/vehicle motion.

Intrusion detection has a little more fine adjustment to it, the initial set-up is the same in that you draw detection areas but that is the only similarity. After setting detection areas you can then set minimum & maximum detection sizes so that small animals or very large vehicles don't trigger the event, you also can adjust sensitivity and threshold which are explained below:

Sensitivity
Sensitivity stands for the percentage of the body part of an acceptable target that enters the
Detection region. Sensitivity = 100 - S1/ST × 100. S1 stands for the target body part that goes
across the Detection region. ST stands for the complete target body. The higher the value of
Sensitivity is, the more easily the alarm can be triggered.

Threshold
Threshold stands for the threshold for the time of the object loitering in the region. If the
time that one object stays exceeds the threshold, the alarm is triggered. The larger the value
of the threshold is, the longer the alarm triggering time is.

So for example, if you had an intrusion detection area set on your driveway and it stretched slightly onto the pavement at the end of your drive, you could set the threshold longer so members of the public just walking past quickly don't trigger the event.
Dear Dan,
Thanks for your answer.
I understand intrusion detection settings but what about when you have to check a long passageway?
I mean if you put a person near the end of the passageway, and measure him as the minimum size (in order the alarm to work when detects someone in the specific point), this minimum size will trigger the alarm if for example a cat go through from the front point of the passageway, because probably it will be larger from the minimum size that I have set. Am I right?
If yes what is the solution for this kind of situations?
I will attach a photo of one of my cameras to understand what am I talking about.
(Sorry about the quality… is from substream)
 

Attachments

  • 4EF2215C-3041-4E07-B3A4-51A464A7E994.jpeg
    4EF2215C-3041-4E07-B3A4-51A464A7E994.jpeg
    551.9 KB · Views: 193
I think it is unlikely that the minimum size of a human at the far end of the wall on the left of the image would be as small as a cat/dog in the foreground (I wouldn't set a smart event any further than the end of the wall because the image quality is poor at that distance and this will impact event accuracy), but even if it was, intrusion detection still supports the Human/Vehicle filtering so with either or both of these enabled the event should be far less likely to trigger on animals.
 
I think it is unlikely that the minimum size of a human at the far end of the wall on the left of the image would be as small as a cat/dog in the foreground (I wouldn't set a smart event any further than the end of the wall because the image quality is poor at that distance and this will impact event accuracy), but even if it was, intrusion detection still supports the Human/Vehicle filtering so with either or both of these enabled the event should be far less likely to trigger on animals.
Thanks for your help Dan.
So from your experience how many meters is good point to end an intrusion area for these cameras (2.8mm lens).
I’m asking that because all of my cameras are “watching” large point of fields areas.
I will attach a day photo of the same area to understand the length. The end of the wall on the left side is about 30meters from the camera. The area above reaches 80-90meters. So in order to have a measurement base to set the intrusion area of each camera and work right, how many meters I have to calculate as the max point for each camera?
Thank you
 

Attachments

  • AA4A4849-06DD-4C5C-9820-4BD761C27D9C.jpeg
    AA4A4849-06DD-4C5C-9820-4BD761C27D9C.jpeg
    505.9 KB · Views: 183
Hi @alexdel

A basic rule, if you are wanting reliable event triggers day & night, is not to go far beyond the IR range of your cameras (e.g. 60m for 2T87G2 and 30/40m for the 2043)

For the 2T87G2 you can also take into account its DORI stats which can be found on the datasheet, see a screenshot below:
Screenshot 2022-09-21 at 15.14.09.png


DORI stands for Detection, Observation, Recognition, Identify and these figures give you a rough idea of the level of detail you will get at different distances, more info on DORI can be found HERE

You might be able to get events to trigger beyond the Observation distance but as the detail/resolution is reduced the accuracy of the Human/vehicle feature will be impacted and you will probably get more false detection, so we wouldn't set the event zone any further than 40m from the 2T87G2 camera (and 20-25m for the 2043G2).
 
Hi @alexdel

A basic rule, if you are wanting reliable event triggers day & night, is not to go far beyond the IR range of your cameras (e.g. 60m for 2T87G2 and 30/40m for the 2043)

For the 2T87G2 you can also take into account its DORI stats which can be found on the datasheet, see a screenshot below:
View attachment 8215


DORI stands for Detection, Observation, Recognition, Identify and these figures give you a rough idea of the level of detail you will get at different distances, more info on DORI can be found HERE

You might be able to get events to trigger beyond the Observation distance but as the detail/resolution is reduced the accuracy of the Human/vehicle feature will be impacted and you will probably get more false detection, so we wouldn't set the event zone any further than 40m from the 2T87G2 camera (and 20-25m for the 2043G2).
Dear Dan, Thank you so much for the time you spend to help me. I would like to ask you a couple more things and when you have time please give me your knowledge.

1) When I check the intrusion detection box of a camera do I have necessarily to draw a box or if I don’t the intrusion zone covers all the view of my camera?

2) My equipment is a DS-7732NXI-I4/S NVR. 10 of my cameras are DS-2CDT87G2-L, 6 DS-2CD2043G2-IU and 3 DS-2CD2T45G0P-I.
I want to enable intrusion detection to all of my DS-2CDT87G2-L cameras (because there are for peripheral protection) and 2 of the DS-2CD2043G2.
But I face this issue.
When I enable the intrusion smart event and draw an area in some cameras when I press apply button it gives me a “parameter error” message. Also when this happens it disorganise the intrusion settings of some of my other cameras. For example by unchecking the vehicle and human check box by itself on the intrusion settings of other cameras. Can you help me on what this parameter error means and why this affects the settings of my other cameras?

And last…
3) I would like one of my DS-2CDT87G2-L camera to capture pictures when a vehicle gets in a pre defined area. NOT to send me alarm or notify me by anyway. Just to capture pictures in order I can watch later in a day the vehicles that have been in this area. But ideally I would like to have the option to see those pictures not only directly from my NVR but also from far away via web browser interface or IVMS-4200 or Hik-Connect (I think it’s not possible via hik-connect but I’m not sure).
So I would like to tell me which is the best way to do that.

Thank you again for your time
 
Hi @alexdel

1) You can draw the intrusion area to cover the whole scene, but we have never tested this and I think there is a good chance it will impact the event accuracy because as explained earlier intrusion sensitivity operates based on the percentage of a body entering the area. If you push the detection area to the edges of the image the camera might struggle to understand when something has entered the detection area because as soon as anything enters the camera's FoV it is already in the detection area.

2) What firmware versions are you running for the NVR & cameras? Have you tried enabling the events from the NVR menus and then just adjusting the settings not available in the NVR menus by directly logging into the cameras? Are you making sure to click 'save' after adjusting any settings?

3) This video explains the basics of how you can use the Human/Vehicle filtering to view stills/clips of vehicle event triggers, if you don't have any linkage action enabled then the events should still trigger but you won't receive any notification. The Human/Vehicle filtering can be done via direct monitor or iVMS-4200. (The linked video is a few years old, you will now find the event settings under 'Smart Analysis' on the direct monitor)
 
Hi @alexdel

1) You can draw the intrusion area to cover the whole scene, but we have never tested this and I think there is a good chance it will impact the event accuracy because as explained earlier intrusion sensitivity operates based on the percentage of a body entering the area. If you push the detection area to the edges of the image the camera might struggle to understand when something has entered the detection area because as soon as anything enters the camera's FoV it is already in the detection area.

2) What firmware versions are you running for the NVR & cameras? Have you tried enabling the events from the NVR menus and then just adjusting the settings not available in the NVR menus by directly logging into the cameras? Are you making sure to click 'save' after adjusting any settings?

3) This video explains the basics of how you can use the Human/Vehicle filtering to view stills/clips of vehicle event triggers, if you don't have any linkage action enabled then the events should still trigger but you won't receive any notification. The Human/Vehicle filtering can be done via direct monitor or iVMS-4200. (The linked video is a few years old, you will now find the event settings under 'Smart Analysis' on the direct monitor)
Dear Dan,

1) I was just asking if it is necessary to draw a box. For example if I check the intrusion detection check box and don’t draw any area will the intrusion detection work at all or it is necessary to draw a box (regardless the coverage area) in order to work?

2) I will attach 2 photos. One of the error message and one of my NVR firmware.
But I didn’t understand what you are saying.
I do all my settings from my NVR menu and not directly from each camera.
Also are there any more settings if I login directly to the camera that are not available in the NVR menu?

3) I will check the video. But as I understand from what you are saying, the only available way to check my stored pictures from my NVR is with IVMS and not from web browser. Is that right?
 

Attachments

  • 609B3BE5-FEC7-4CFF-921D-41F0C8C13ED9.jpeg
    609B3BE5-FEC7-4CFF-921D-41F0C8C13ED9.jpeg
    1.1 MB · Views: 159
  • 83A867E1-9E06-4787-9456-06B5C3FA5476.jpeg
    83A867E1-9E06-4787-9456-06B5C3FA5476.jpeg
    1.9 MB · Views: 156
1) Yes, a box needs to be drawn for intrusion detection to work

2) The latest firmware for your NVR is V4.61.020, try updating to the STD version of this firmware and see if that removes the error message. On the direct monitor for the NVR, it looks like you do have access to all the event settings (I thought you might have been accessing with a browser, in the NVR browser access, not all settings are accessible).

3) You don't have access to the human/vehicle filtering feature in browser playback, it is not currently possible to see captures of vehicle event triggers in the browser access.
 
1) Yes, a box needs to be drawn for intrusion detection to work

2) The latest firmware for your NVR is V4.61.020, try updating to the STD version of this firmware and see if that removes the error message. On the direct monitor for the NVR, it looks like you do have access to all the event settings (I thought you might have been accessing with a browser, in the NVR browser access, not all settings are accessible).

3) You don't have access to the human/vehicle filtering feature in browser playback, it is not currently possible to see captures of vehicle event triggers in the browser access.
Dear Dan,
About the error message. Is it possible that the NVR has limit on how many cameras you can set intrusion detection? Or is it an issue if for example I have in one camera both intrusion and line crossing enable?
Also get in mind that I have enable all of my cameras the basic event motion detection.
 
Hi @alexdel

It could be an issue of multiple events being set, we have seen issues with this on certain models in the past.

I would try disabling line crossing and then try setting intrusion detection to see if that removes the error message.
 
Hi @alexdel

It could be an issue of multiple events being set, we have seen issues with this on certain models in the past.

I would try disabling line crossing and then try setting intrusion detection to see if that removes the error message.
What about my first question? If there is a limit on the NVR on how many cameras can I enable intrusion detection?
Because I read this on the description. Up to 4ch perimeter protection.
 
Back
Top