Difference between sensibility and threshold

netcamstudio

#1

Hi !

What is the difference between sensibility and threshold ?

I understand threshold is the % of image that need to change to trigger motion but I don’t know what sensibility is.

Regards,

JB


Motion Detection Problem
#2

Hi JB!
Excellent question! To know how Threshold and Sensibility works is very important to make motion detection a success or a failure. So here comes a short story of them both. Maybe a little bit longer that I expected, but I found it interesting to write about ;). If you do not understand anything please, reply that it is crap and I will try again ;).

As the name reveals it is about detection of an object in motion. So the system need first to verify that it is an object in the field of view of the camera and second if the object is moving. Since it´s all mathematics the system need two parameters; let´s name them threshold value for the detection of an object and sensibility value for the detection of movement.

Threshold value
As you mentioned the threshold value is a number in % and is the amount of pixels that is covered by the object relative the total amount of pixels in the picture. Let’s say that a person cover 15% of the total amount of pixels. Then the threshold value is 15%. For the system to detect that person the threshold level must be set to a lower value than 15%, let´s say 10%. Now, if something covers more than 10% of the pixels in the field of view the system will acknowledge that as an object of interest. But, that information will solely not trigger the system for motion detection… Therefore, we need a second parameter called Sensibility value.

Sensibility value
Sensibility is a parameter connected to if the object is moving slowly or fast. If the object is moving at all or if it is moving slowly or fast is determined from more mathematical calculations during a certain time interval. Let´s say that at time=1 the calculation gives the first threshold value. A specific time later at time=2 the second calculation gives the next threshold value and after time=3 it gives the third threshold value. If all these threshold values are the same the object is not moving and the system will not generate a motion trigg to start a recording of a video. But, if it is a difference between the threshold values something is happening with the object and it gets interesting. If the differences between the threshold values are small the object is probably moving slow. If the difference is big the object is probably moving fast. Let´s say that the system calculates a sensibility value that is between 0 and 100. A value of 0 means no change in motion, completely still. A value of 100 means t that the object is moving very fast. In Netcam system the Sensibility can be set to; very high – normal – very low. If you want to detect an object that moves very slowly you need to set the Sensibility to very high. Small changes in the sensibility value should be recognized as important and tell the system that here is something important going on. If the object is moving fast the sensibility value is high and it will also be recognized by the system as important. Well, why not always set the Sensibility to very high because then we will never miss a moving object? That is the million dollar question. If you set the Sensibility to very high it will detect everything and you will have a lot of false alarms! If you set the Sensibility to very low you will most likely never get a false alarm, but you will never get the moving object of interest either. Setting the Sensibility correct needs quite often some testing since it is very dependent of the situation.

Summary
To summarize setting the Threshold Level to a correct value let the system know that the object is of interest. Setting the Sensibility correct let the system know that the object of interest is moving. So the slower the object is moving the higher needs the Sensibility to be. These two parameters together trigger the motion detection system and it starts to record a video.

This was in very general terms how this works. There are many different mathematical algorithms depending on what object you want to detect and so on. In high-end application they also use Artificial Intelligence. Finally, if you know what object you want to detect and also a little bit how fast it will move the more likely it is that you get very good results from the system. But, usually some testing is needed to get it right.

From this comes also an interesting result why the cams of today that have megapixel sensors need rather powerful computers for motion detection, but that is another story ;).

-Henrik


#3

Thank you very much !

Now I am searching a setting like “motion duration”. I tried “trigger duration” but it doesn’t do what I want. I would like to trigger a “motion event” only if the “event” is still in motion after X seconds since the first motion was detected. Is it possible ? Maybe this feature is not available yet I can’t find it.

Regards,

JB


#4

The Trigger Duration sets how long time a video is recorded after a motion trigg. If you get a motion trigg during that time it will continue to record with another Trigger Duration time.

By using a Rule you can start an event at the same time as the motion trigger. But, in NCS there are not today support for a delayed start of a second event. However, with an external home security systems it would be possible. There are people here on the forum that sends data from NCS to their home security systems.

-Henrik


#5

Hi,

I am just trying to get my settings all working well and this has been really helpful.

When working with the threshold value does this consider only the area that is selected for motion detection or does it consider the whole picture? Ie if I had only 10 percent of the picture enabled for motion detection would the threshold relate to a percentage of this 10 percent? (Not the whole capture image)

Cheers,


#6

Hi!
Very interesting question! That is a programming issue and I do not know that. Your question is actually if the Detection Zone affects the Threshold value. I have to direct this question to our developer @Steve to answer.
-Henrik


#7

It still considers the full picture because masking is only forcing and area to be black so that it’s not being detected. so it’s 10% of the full picture not 10% of the selected area.

Anyway the best is to display the gauge / meter while configuring to be able to compare what is reported compared to the target you have or to rely on blob detection which seems more much efficient.


#8

On Threshold…

My views are broken into a grid of 16x24 which is 384 squares. Setting a threshold of 1 should exclude an object smaller than 3.8 squares. I don’t see that the threshold correlates to a percent of the view.

With a threshold set to 2 (should exclude an object smaller than 7 squares) - I am still recording on very small objects moving in the view. My cat will set off my basement camera to record every time. Or a rabbit (smaller than 1 square) still sets off another camera; same camera at nights will be activated by an insect which even up close is smaller than 1%.

Insects and tree shadows are the main issues I have at this time with false alarms.


#9

Hi!
It’s about the number of pixels of the image sensor in the camera.
-henrik


#10

Same here! Henrik’s explanation was great, but it just doesn’t seem to mesh well with the actual threshold settings in Netcam Studio. I have my threshold rather high to try to remove a bunch of my false readings, but still get recordings of insects and sun/cloud shadow changes on the ground that drive me crazy.

If only NCS could detect and differentiate between objects, people, etc. rather than just pure movement…


#11

Hi Gunn!
I certainly agree that it is a drag. I have the same problems especially with shadows that are dancing on the walls. It would be very nice to have a software that could do what you propose. There are software’s or rather systems for that, but they comes with a completely different price tag. Have you tested using another Motion Algorithm.

Problems with insects are usually when the IR is on. To get away from that people usually set up a separate IR source away from the camera. Insects do like the IR since it is very bright for them and also generates heat.

Sorry, but it´s not that much more I can suggest than the above.
-Henrik


#12

I realized about a week ago that I could turn off the IR on the outside camera and get a much cleaner night time image with the ambient light. I turned up the brightness on the camera a bit to compensate too. I realized the IR was causing too much visual noise from insects, dust, and precipitation and false motion events. I hadn’t thought about using a separate IR light-- thanks for the idea! I also always wondered why the camera was an insect magnet-- I am constantly clearing off spiders, flies, etc. Thanks for the tip that insects are attracted to the IR-- it makes sense now!


Motion Detection - Video saved without motion detected
#13

Excellent and glad to help!
-Henrik


#14

I too am struggling with insect false alarms. Also with alarms if there is a glitch in the image and all pixels shift, for example. What would be ideal for me is to require 1 to 2 seconds of motion to set alarm, and then save clip starting -5 seconds back from buffer. I would like to ignore drastic motion like entire image shifting by 1 pixel or a bug flying across image in less than a second. I think requiring 1 to 2 seconds of motion – the same motion target (so is that a blob?) would eliminate most of my false alarms.

Or some other method or reducing “noise” in the image in the form of bugs and camera glitches. Maybe cut everything moving beyond a certain velocity because nothing out many feet from the camera can move as fast as the bugs flying across the camera very close.

Some software I have tested has “ignore motion of less than XX milliseconds” setting. Is there something like this in NCS? Or another setting in other software that I have seen is minimum/maximum object size. Setting a max object size would remove camera pixel shifts (just noticed not all are glitches – some are vibrations from doors opening/closing).

What would be my best approach using settings available to me?


#15

What will the setting “Threshold 0” do?


#16

I noticed that setting it to 0 was “automatic” so I thought maybe some logic there would help but still lots of false alarms. Hundreds last night. Does it improve over time? Maybe it’ll be better tonight? Also not sure which method I should use for detection – seems like maybe “blob” (is this an object like person/animal/vehicle?) but the manual only says “The algorithm counts moving objects and highlights them with red rectangle.” I’m less interested in the count of moving objects… I need to identify a single object of interest that is moving.

I have had it set to background.

It is a little confusing as to why you can choose a display (algorithm?) different than the “motion algorithm” itself. I would think that the purpose of the display would be to help you understand what the “motion algorithm” is doing so shouldn’t it always be set to correspond?


#17

CA_Tallguy, do you use a high or a low resolution camera? I found that going from a very cheap 640x480 to a good 1920x1080 reduced the sensitivity. I actually kind of have the opposite problem, motion is not picking up smaller things, like the neighbors cat pissing on my stairs… I guess that an object appears larger when it’s on a picture with less pixels.


#18

Usually, higher resolution cams cover a larger viewing angle which is the big effect.


#19

What will the setting “Threshold 0” do?


#20

What will the setting “Threshold 0” do? Distance and high resolution makes every moving objects a low pixel count object, so I guess I should use a low Threshold value, but what is the difference in “0” and “1”?