Understanding ZoneMinder's Zoning system for Dummies - Revision history

Gerdesj: Fix forum links

2021-07-06T15:25:33Z

Fix forum links

← Older revision		Revision as of 11:25, 6 July 2021
Line 5:		Line 5:
	ZoneMinder has a flexible (albeit hard to easily configure) zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [https://zoneminder.readthedocs.io/en/stable/userguide/definezone.html here]. However, you will see gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.		ZoneMinder has a flexible (albeit hard to easily configure) zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [https://zoneminder.readthedocs.io/en/stable/userguide/definezone.html here]. However, you will see gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.

	So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://~~www~~.zoneminder.com~~/forums~~/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://~~www~~.zoneminder.com~~/forums~~/viewtopic.php?f=11&t=22239&p=84476#p84476 here])		So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://forums.zoneminder.com/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://forums.zoneminder.com/viewtopic.php?f=11&t=22239&p=84476#p84476 here])

	===Real object detection===		===Real object detection===

Gerdesj: Fix documentation link and some minor spellings

2021-07-06T15:19:57Z

Fix documentation link and some minor spellings

← Older revision		Revision as of 11:19, 6 July 2021
Line 3:		Line 3:
	===Background===		===Background===

	ZoneMinder has a flexible (albeit hard to easily configure) zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [~~http~~://~~www~~.~~zoneminder~~.~~com~~/~~wiki~~/~~index~~.~~php/Defining_Zones~~ here]. However, you will see		ZoneMinder has a flexible (albeit hard to easily configure) zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [https://zoneminder.readthedocs.io/en/stable/userguide/definezone.html here]. However, you will see gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.
	gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.

	So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])		So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])
Line 13:		Line 12:

	===Some concepts===		===Some concepts===

	Let's take a look at this area below. Lets suppose you want to trigger motion if someone tries to break into your basement. Does it make sense you monitor the full area (pillars/walls/floor)? Probably not. If someone were to break in, they'd break in from some door, some window, or maybe break in from upstairs and climb down the stairs. So doesn't it make more sense to monitor these areas specifically? I think so. So the first 'common sense' logic is delete the default zone that ZM creates for each monitor (which is called All). Monitoring every part of your image may make sense if you are monitoring and outdoor lawn, for example. Not here.		Let's take a look at this area below. Lets suppose you want to trigger motion if someone tries to break into your basement. Does it make sense you monitor the full area (pillars/walls/floor)? Probably not. If someone were to break in, they'd break in from some door, some window, or maybe break in from upstairs and climb down the stairs. So doesn't it make more sense to monitor these areas specifically? I think so. So the first 'common sense' logic is delete the default zone that ZM creates for each monitor (which is called All). Monitoring every part of your image may make sense if you are monitoring and outdoor lawn, for example. Not here.

Line 20:		Line 18:

	=== Defining the zone areas ===		=== Defining the zone areas ===
	So given the explanation above, how about we define zones where motion matters? Any zone you define as "active" is what ZoneMinder will analyze for motion. Ignore the 'preclusive for now'. So ~~lets~~ look at the image below. I've defined polygons around places that are the "entry points"		So given the explanation above, how about we define zones where motion matters? Any zone you define as "active" is what ZoneMinder will analyze for motion. Ignore the 'preclusive for now'. So let's look at the image below. I've defined polygons around places that are the "entry points"

	[[File:With_zones.jpg]]		[[File:With_zones.jpg]]
Line 31:		Line 29:

	==== Core Concepts ====		==== Core Concepts ====
	The ZM ~~wiki~~ I pointed to earlier does a great job of explaining different methods. At the cost of repeating what has already been said, ~~its~~ important to note:		The ZM documentation I pointed to earlier does a great job of explaining different methods. At the cost of repeating what has already been said, it's important to note:

	* ZM does NOT understand objects. It only understands pixel colors. So if you are monitoring a camera that is producing a 1280x960 32bit color depth image, as far as ZM is concerned, it is getting an array of 128096032 bits of data to analyze and compare a previous frame and based on 'color differences' between frame X-1 and X along with some algorithms it applies in addition to color differences, it tries and guesses if objects (in ZM speak, a specific pattern of pixels) have come up that were not there previously.		* ZM does NOT understand objects. It only understands pixel colors. So if you are monitoring a camera that is producing a 1280x960 32bit color depth image, as far as ZM is concerned, it is getting an array of 128096032 bits of data to analyze and compare a previous frame and based on 'color differences' between frame X-1 and X along with some algorithms it applies in addition to color differences, it tries and guesses if objects (in ZM speak, a specific pattern of pixels) have come up that were not there previously.
Line 50:		Line 48:

	=====Filtered Pixels=====		=====Filtered Pixels=====
	Now let's assume we used Filtered pixels and set it to 2x2 pixels. ~~The~~ in addition to computing the alarmed pixels (A+B+C+D), it will also count how many of these sets have at least 2 pixels around them that are also alarmed pixels. This will result in B+C+D (set A will be discarded as they don't have any pixels surrounded by at least 2 pixels that have changed color from the prev. frame)		Now let's assume we used Filtered pixels and set it to 2x2 pixels. Then in addition to computing the alarmed pixels (A+B+C+D), it will also count how many of these sets have at least 2 pixels around them that are also alarmed pixels. This will result in B+C+D (set A will be discarded as they don't have any pixels surrounded by at least 2 pixels that have changed color from the prev. frame)

	=====Blob Pixels=====		=====Blob Pixels=====
	Now ~~lets~~ assume we used Blob and said a blob needs to be at least 10 pixels. Then what it will do is based on the set computed by Filtered pixels, which is B+C+D it will look for contiguous blobs of 10 pixels and that only means D		Now let's assume we used Blob and said a blob needs to be at least 10 pixels. Then what it will do is based on the set computed by Filtered pixels, which is B+C+D it will look for contiguous blobs of 10 pixels and that only means D

	So, in Alarmed pixels any of A, B, C or D would raise an alarm		So, in Alarmed pixels any of A, B, C or D would raise an alarm
Line 59:		Line 57:
	In blob mode only D would raise an alarm		In blob mode only D would raise an alarm

	Okay, that was a simple explanation. And I did not cover more details on min/max. But I hope you get the core idea..		Okay, that was a simple explanation. And I did not cover more details on min/max. But I hope you get the core idea.

	===Got the theory. Let's get back to your basement image===		===Got the theory. Let's get back to your basement image===
Line 75:		Line 73:

	=====Using percents=====		=====Using percents=====
	(As of 2019, I recommend pixels, but I've kept this here for those who prefer ~~percents~~)		(As of 2019, I recommend pixels, but I've kept this here for those who prefer percentages)

	It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:		It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:
	[[File:Of_men_and_animals.jpg]]		[[File:Of_men_and_animals.jpg]]

	* '''Zones and Depths:''' One of the ~~learning~~ for me was ~~that~~ don't make zones cover a lot of depth as far as the camera lens goes. Why? Well take a look at the image above. As a person or a pet walks away or towards the camera, it gets significantly bigger or smaller (right, that was not rocket science, but sometimes commonsense 101 helps)		* '''Zones and Depths:''' One of the lessons for me was don't make zones cover a lot of depth as far as the camera lens goes. Why? Well take a look at the image above. As a person or a pet walks away or towards the camera, it gets significantly bigger or smaller (right, that was not rocket science, but sometimes commonsense 101 helps)

	* One of the advantages of using ~~percents~~ is that if you were to change your zone size, the percent settings will likely remain consistent (as they will be percentages of your zone area). In pixel mode, any change to the zone polygon will require updating the other values		* One of the advantages of using percentages is that if you were to change your zone size, the percent settings will likely remain consistent (as they will be percentages of your zone area). In pixel mode, any change to the zone polygon will require updating the other values

	* This is what my zone definition for the "stairs area" looks like, using percents:		* This is what my zone definition for the "stairs area" looks like, using percents:

Asker at 12:41, 30 June 2019

2019-06-30T12:41:18Z

← Older revision		Revision as of 08:41, 30 June 2019
Line 95:		Line 95:
	* You should also look at filling in a "max" area - this is to avoid sudden "blast of light changes" that raise an alarm. For example, if most of my zone pixels changed, it is unlikely its an object of interest. It is more likely the sun just blasted light on my driveway or a big shadow situation occurred. ZM helpfully tells you your zone area- in this case 92271. You could choose to put in a pixel value that is some large fraction of this value (maybe 90% of it = 83,000 or simply draw another rectangle to represent the largest object that is of interest to you and use that area value)		* You should also look at filling in a "max" area - this is to avoid sudden "blast of light changes" that raise an alarm. For example, if most of my zone pixels changed, it is unlikely its an object of interest. It is more likely the sun just blasted light on my driveway or a big shadow situation occurred. ZM helpfully tells you your zone area- in this case 92271. You could choose to put in a pixel value that is some large fraction of this value (maybe 90% of it = 83,000 or simply draw another rectangle to represent the largest object that is of interest to you and use that area value)

	'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at [https://github.com/pliablepixels/zmeventnotification zmeventnotification], the machine learning based alarm detection extension to ZM~~'''~~		'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit.''' You'll never get it perfect. For real object detection as an 'add-on' to motion, look at [https://github.com/pliablepixels/zmeventnotification zmeventnotification], the machine learning based alarm detection extension to ZM

	[[File:pixel_area.jpg\|800px ]]		[[File:pixel_area.jpg\|800px ]]

Asker at 12:39, 30 June 2019

2019-06-30T12:39:39Z

← Older revision		Revision as of 08:39, 30 June 2019
Line 3:		Line 3:
	===Background===		===Background===

	ZoneMinder has a ~~powerful~~ zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [http://www.zoneminder.com/wiki/index.php/Defining_Zones here]. However, you will see		ZoneMinder has a flexible (albeit hard to easily configure) zone detection system using which you can modify how sensitive, precise, accurate your motion alarms are. The official ZM documentation does a good job of describing all the concepts [http://www.zoneminder.com/wiki/index.php/Defining_Zones here]. However, you will see
	gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.		gobs of posts in the forum about people complaining that ZM logs all sorts of events (ahem, as did I), ZM's detection is rubbish and in-camera is better (ahem, as did I) and what not. But once you get the concept, its incredibly powerful.

	So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])		So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])

			===Real object detection===
			People often ask if ZM supports "object detection". The ''core'' ZM engine only detects change in pixels (motion), which is what this article is about. That being said, incase you did not know, ZoneMinder now has support for person, object, face recognition. So while you can and should optimize your zones, if your real interest is person detection, you can do that using my [https://github.com/pliablepixels/zmeventnotification event server]. Note that object detection works ''after'' ZM detects motion change, so this article applies either way.

	~~===Updated on Apr 2019===~~

	Incase you did not know, ZoneMinder now has support for person, object, face recognition. So while you can and should optimize your zones, if your real interest is person detection, you can do that using my [https://github.com/pliablepixels/zmeventnotification event server]

	===Some concepts===		===Some concepts===

	Let's take a look at this area below. Lets suppose you want to trigger motion if someone tries to break into your basement. Does it make sense you monitor the full area (pillars/walls/floor)? Probably not. If someone were to break in, they'd break in from some door, some window, or maybe break in from upstairs and climb down the stairs. So doesn't it make more sense to monitor these areas specifically? I think so. So the first 'common sense' logic is delete the default zone that ZM creates for each monitor (which is called All). Monitoring every part of your image may make sense if you are monitoring and outdoor lawn, for example. Not here		Let's take a look at this area below. Lets suppose you want to trigger motion if someone tries to break into your basement. Does it make sense you monitor the full area (pillars/walls/floor)? Probably not. If someone were to break in, they'd break in from some door, some window, or maybe break in from upstairs and climb down the stairs. So doesn't it make more sense to monitor these areas specifically? I think so. So the first 'common sense' logic is delete the default zone that ZM creates for each monitor (which is called All). Monitoring every part of your image may make sense if you are monitoring and outdoor lawn, for example. Not here.

	[[File:Nph-zms.jpeg]]		[[File:Nph-zms.jpeg]]
Line 62:		Line 61:
	Okay, that was a simple explanation. And I did not cover more details on min/max. But I hope you get the core idea..		Okay, that was a simple explanation. And I did not cover more details on min/max. But I hope you get the core idea..

	===~~Yawn. Forget~~ the theory. Let's get back to your basement image===		===Got the theory. Let's get back to your basement image===

	Okay, back to my basement and my 3 zones.		Okay, back to my basement and my 3 zones.
Line 76:		Line 75:

	=====Using percents=====		=====Using percents=====
			(As of 2019, I recommend pixels, but I've kept this here for those who prefer percents)

	It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:		It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:
	[[File:Of_men_and_animals.jpg]]		[[File:Of_men_and_animals.jpg]]
Line 81:		Line 82:
	* '''Zones and Depths:''' One of the learning for me was that don't make zones cover a lot of depth as far as the camera lens goes. Why? Well take a look at the image above. As a person or a pet walks away or towards the camera, it gets significantly bigger or smaller (right, that was not rocket science, but sometimes commonsense 101 helps)		* '''Zones and Depths:''' One of the learning for me was that don't make zones cover a lot of depth as far as the camera lens goes. Why? Well take a look at the image above. As a person or a pet walks away or towards the camera, it gets significantly bigger or smaller (right, that was not rocket science, but sometimes commonsense 101 helps)

	* ~~'''Percent vs. Pixels:''' Back to my previous point -~~if you ~~specify alarmed areas in pixels~~, ~~in addition to being able to visualize better~~ (~~at least for me) the added advantage is the setting scales~~ as ~~zone sizes changes. For example the %~~ of ~~space the person takes in my far end "Door"~~ zone ~~is quite similar to the % of space the person takes in my couch~~ area ~~(in other words~~, ~~both~~ the zone ~~and person have scaled up in size, so relative % is still~~ the ~~same)~~		* One of the advantages of using percents is that if you were to change your zone size, the percent settings will likely remain consistent (as they will be percentages of your zone area). In pixel mode, any change to the zone polygon will require updating the other values

	* ~~So having said all of this, this~~ is what my zone definition for the "stairs area" looks like:		* This is what my zone definition for the "stairs area" looks like, using percents:
	[[File:Stairs foyer.png]]		[[File:Stairs foyer.png]]

Line 94:		Line 95:
	* You should also look at filling in a "max" area - this is to avoid sudden "blast of light changes" that raise an alarm. For example, if most of my zone pixels changed, it is unlikely its an object of interest. It is more likely the sun just blasted light on my driveway or a big shadow situation occurred. ZM helpfully tells you your zone area- in this case 92271. You could choose to put in a pixel value that is some large fraction of this value (maybe 90% of it = 83,000 or simply draw another rectangle to represent the largest object that is of interest to you and use that area value)		* You should also look at filling in a "max" area - this is to avoid sudden "blast of light changes" that raise an alarm. For example, if most of my zone pixels changed, it is unlikely its an object of interest. It is more likely the sun just blasted light on my driveway or a big shadow situation occurred. ZM helpfully tells you your zone area- in this case 92271. You could choose to put in a pixel value that is some large fraction of this value (maybe 90% of it = 83,000 or simply draw another rectangle to represent the largest object that is of interest to you and use that area value)

	'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''		'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at [https://github.com/pliablepixels/zmeventnotification zmeventnotification], the machine learning based alarm detection extension to ZM'''

	[[File:pixel_area.jpg\|800px ]]		[[File:pixel_area.jpg\|800px ]]
Line 121:		Line 122:
	#''' Raise Alarm!'''		#''' Raise Alarm!'''

	Note the percentages I describe above seem lesser than the visual percentages when you look at the image. First, my image is not to scale. Second, I guess I could increase the percentages, but these values are working quite well for me, but in reality I guess I could easily increase, say, 6% to 10%. In general, I'd recommend using the least value that is not triggered by small things - keep raising it till you stop getting 90% of spurious alarms, but no further. If you try and eliminate all spurious events, then its likely you will also miss some valid motion (that's a non scientific anecdotal statement, so feel free to challenge me here)

Asker at 12:30, 30 June 2019

2019-06-30T12:30:40Z

← Older revision		Revision as of 08:30, 30 June 2019
Line 123:		Line 123:
	Note the percentages I describe above seem lesser than the visual percentages when you look at the image. First, my image is not to scale. Second, I guess I could increase the percentages, but these values are working quite well for me, but in reality I guess I could easily increase, say, 6% to 10%. In general, I'd recommend using the least value that is not triggered by small things - keep raising it till you stop getting 90% of spurious alarms, but no further. If you try and eliminate all spurious events, then its likely you will also miss some valid motion (that's a non scientific anecdotal statement, so feel free to challenge me here)		Note the percentages I describe above seem lesser than the visual percentages when you look at the image. First, my image is not to scale. Second, I guess I could increase the percentages, but these values are working quite well for me, but in reality I guess I could easily increase, say, 6% to 10%. In general, I'd recommend using the least value that is not triggered by small things - keep raising it till you stop getting 90% of spurious alarms, but no further. If you try and eliminate all spurious events, then its likely you will also miss some valid motion (that's a non scientific anecdotal statement, so feel free to challenge me here)

	~~====Door Zone====~~
	~~Here is my zone setting for the door (Edit note: the max alarmed/filtered area was changed to 70% - the image shows 90%)~~

	~~[[File:Door.png]]~~

	The logic is similar to the foyer area - plus, I've used the overload frame count -- basically the sun blasts through the blinds and in those cases the % change is very large (around 75+%) which triggers my max % spec and in that case, I tell zoneminder to ignore the next 2 frames so that this excessive lighting condition does not raise lots of alarms. BTW, this is one zone where I still get 4-5 false events in a day. I am fine with it and prefer to live with it.

	~~====Ah, I get it. But hey what's that preclusive zone in your image at the start of this artlce?====~~
	~~Ah, ZM allows us to specify a zone where should an alarm occur, it suppresses alarms in that zone and all other zones in the monitor. I've set it to cover my recess lights.~~
	~~So if someone turns our lights on, I won't get an alarm. Does it work? Yep pretty well. Does it always work? Mostly. Heh heh.~~

	===Some other optimizations===		===Some other optimizations===
	There are other optimizations you can do. Another setting I found useful is to use "Alarm Frame Count". This is not part of Zones, but part of your monitor definition (Buffers tab). It specifies the minimum amount of alarm frames to trigger an alarm. Most of my cameras are on Wifi and my basement is farthest away from my AP, so I occasionally get smearing. Setting this to 3 reduces almost all smears raising an alarm		There are other optimizations you can do:
			* Another setting I found useful is to use "Alarm Frame Count". This is not part of Zones, but part of your monitor definition (Buffers tab). It specifies the minimum amount of alarm frames to trigger an alarm. Most of my cameras are on Wifi and my basement is farthest away from my AP, so I occasionally get smearing. Setting this to 3 reduces almost all smears raising an alarm
	~~===You're a dork. None of this mumbo jumbo worked for me.~~ I ~~want my money back===~~		* As I wrote earlier, use maximum values in the zone definition to reduce 'change bursts' (significant pixel coloration changes due to spikes of light/shadows)
	~~Life sucks. Take a break. Drink a glass of wine. Or two. Then get back~~ to ~~your settings. And its possible these don~~'~~t work for you - try different values. Think visually.~~
	~~Oh, and I~~'~~ll need a postage paid envelope~~ to ~~email you a $0 cheque. And a $5 cheque from you to pay for gas to drive to the nearest post box. Labor is free because you did not like my article. There is a lifetime return policy, you see.~~

Asker at 12:28, 30 June 2019

2019-06-30T12:28:38Z

← Older revision		Revision as of 08:28, 30 June 2019
Line 96:		Line 96:
	'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''		'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''

	[[File:pixel_area.jpg]]		[[File:pixel_area.jpg\|800px ]]

Asker at 12:26, 30 June 2019

2019-06-30T12:26:20Z

← Older revision		Revision as of 08:26, 30 June 2019
Line 96:		Line 96:
	'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''		'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''

	[[File:~~pixel area~~.~~png~~]]		[[File:pixel_area.jpg]]

Asker at 12:21, 30 June 2019

2019-06-30T12:21:47Z

← Older revision		Revision as of 08:21, 30 June 2019
Line 73:		Line 73:
	(Pixels are especially useful if the difference in sizes is small. Example: its not possible to visualize the difference between 10% and 15%)		(Pixels are especially useful if the difference in sizes is small. Example: its not possible to visualize the difference between 10% and 15%)

	====~~Right percents. But what~~ values?====		====Selecting the right values====

			=====Using percents=====
	It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:		It helps to think visually here. Let's go back to the zones I drew of my basement and try and visually place how a person and a pet would look in each zone. Here is a take:
	[[File:Of_men_and_animals.jpg]]		[[File:Of_men_and_animals.jpg]]
Line 83:		Line 84:

	* So having said all of this, this is what my zone definition for the "stairs area" looks like:		* So having said all of this, this is what my zone definition for the "stairs area" looks like:
			[[File:Stairs foyer.png]]


			=====Using pixels=====
			(This is what I ended up using once I got comfortable with zones)
			* This was updated Jun 2019, I don't have that basement camera any more, so I am going to use a different image
			* I usually start with a Min/Max pixel threshold of 35 (that is, consider pixels alarmed only if the color different between the old pixel and the new one is 35 or more)
			* For the min/max alarmed area, I draw a box inside my zone that represents an object size I think is appropriate. In the image below, I am trying to estimate what the size of a person would be in the zone. I use a mac - I simply hit Cmd+Shift+4 - that brings up a screen capture cross hair and as I draw the box, I see its width and height. Once I have a box with the dimension that I think is appropriate, I hit escape and simply fill in Min/Max Area = W*H of that box I used as a visual tool (the image shows a green box, just as a representation). In this case, my box was 42X67 pixels, so I fill in a min area of 2800
			* You should also look at filling in a "max" area - this is to avoid sudden "blast of light changes" that raise an alarm. For example, if most of my zone pixels changed, it is unlikely its an object of interest. It is more likely the sun just blasted light on my driveway or a big shadow situation occurred. ZM helpfully tells you your zone area- in this case 92271. You could choose to put in a pixel value that is some large fraction of this value (maybe 90% of it = 83,000 or simply draw another rectangle to represent the largest object that is of interest to you and use that area value)

	=~~===Stairs Zone====~~		'''Tip: Don't try to get super precise in your first round: Use less aggressive values to begin with - that is, lesser area values. Keep increasing and testing till you reach a good threshold between bogus detections and real detections. Pixel based detections have a limit. You'll never get it perfect. For real object detection, look at zmeventnotification, the machine learning based alarm detection extension to ZM'''

			[[File:pixel area.png]]


			* (Note 1: the image is not to scale, so measurements will not tally if you are trying to draw boxes around the example image above)
			* (Note 2: when you are measuring sizes, make sure you have not reduced browser scale. If you did, obviously, your pixel markings will be wrong because the image will be scaled)
			* I then think about what might be the area inside that rectangle I drew that will have a continuous change in pixels. My guess here was 9x9 which is what I put into the Filter Pixels
			* Finally, since I put a value of 2800 in the alarmed pixel area, obviously Min Alarmed Area > Min Filtered Area > Min Blob Area. So I set Min Filtered Area to around 35% of 2800 = 1000 and blob area to around 90% of filtered area (approx 900), the reason being in my estimate blob area should closely match filtered area of contiguous pixels for humans. This is my guess shrug

	~~[[File:Stairs foyer.png]]~~

	Lets see how ZM will use these values:		Lets see how ZM will use these values:
			'''(While the example below refers to percents, the same logic applies to pixels)'''

	* Use the "blob" method to detect movement (which means first use Alarmed pixels, then use Filtered Pixels, then use Blob)		* Use the "blob" method to detect movement (which means first use Alarmed pixels, then use Filtered Pixels, then use Blob)

Asker: /* Pixels or percents? */

2019-06-29T21:40:07Z

Pixels or percents?

← Older revision		Revision as of 17:40, 29 June 2019
Line 69:		Line 69:

	====Pixels or percents?====		====Pixels or percents?====
	What makes more sense to you? "Raise an alarm if 178 pixels are changed" or "Raise an alarm if more than 20% of my zone has changed?". To me, the latter makes much more sense.		What makes more sense to you? "Raise an alarm if 178 pixels are changed" or "Raise an alarm if more than 20% of my zone has changed?". To some, the latter makes much more sense. However, if you really want more fine grained control, you should use pixels. I used percent when I first started off, but then realized that pixels was more powerful when you are trying to eliminate false alarms.
	So I~~'ve always~~ used ~~percents~~
	~~(Updated in 2019: Given a lot of people seem to refer to this article~~, ~~in some cases, it does make~~ more ~~sense~~ to ~~use pixels,~~ especially if the difference in sizes is small. Example: its not possible to visualize the difference between 10% and 15%)		(Pixels are especially useful if the difference in sizes is small. Example: its not possible to visualize the difference between 10% and 15%)

	====Right percents. But what values?====		====Right percents. But what values?====

Asker at 15:57, 11 April 2019

2019-04-11T15:57:18Z

← Older revision		Revision as of 11:57, 11 April 2019
Line 8:		Line 8:
	So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])		So instead of giving you a theoretical explanation, lets walk through a live use-case. (Credit: user kkrofft helped a lot in me getting a hang of things [http://www.zoneminder.com/forums/viewtopic.php?f=32&t=23024 here]. You should also read his earlier explanation [http://www.zoneminder.com/forums/viewtopic.php?f=11&t=22239&p=84476#p84476 here])


			===Updated on Apr 2019===

			Incase you did not know, ZoneMinder now has support for person, object, face recognition. So while you can and should optimize your zones, if your real interest is person detection, you can do that using my [https://github.com/pliablepixels/zmeventnotification event server]

	===Some concepts===		===Some concepts===