Axis VMD4 Analytics Tested

By IPVM Team, Published on Apr 17, 2018

Axis is now on its 4th generation of video motion detection (VMD), which Axis calls "a free video analytics application." In this generation, Axis emphasizes:

reduce false alarms by analyzing and disregarding normal background motion. Axis False Alarm Filtering provides an additional layer of defense against false alarms by making it possible to ignore disturbing object motion such as headlights, swaying trees and small animals.

We tested VMD4 indoors and outdoors for over two weeks to see how it performed, how it compared to their previous VMD3 app, what triggered false alarms, what did not, and how it integrates (or not) with Avigilon, Exacq, Genetec, Milestone and Network Optix VMSes.

This is the first in a 2018 series of on-board camera analytics tests. Coming up, we plan to release tests of Dahua, Hanwha, and Hikvision analytics, as well.

*******

*******, **** **** ****** very ****** ***** ***** rejection **** ****** ******* such ** ********* ******, lights ******* **/***, ******** monitors, ***. **** *** true *********** *** ** the ***, **** ** settings ******* ********.

********, **** ******** ****** from ******* ********/*******, *** was *********** ** ***** triggers ** ******* *******, such ** ******* **** or **** ***** ** cameras' *****.

***** ******* **** **** an ***** *** ***** be ******* ** ********** by ******** ***** ****** settings ** **** *****. However, ***** ******* *** long-range ********* ****** ********* this ******* *********, ** increasing ******* ****** **** could ********* ********* ** human ******** ** *** far ***** ** ****.

*******, ***** *** *********** has ******** ******** ** VMD3, **** *****, *******, Milestone, *** ******* ***** supporting *** ********, **** support ****** ** ******** for **** ********** *** (and ******* *****) ** it ********* ******** **** level ** *********** ** the ***. 

Test ******* ********

** ******** ******* **** our ******* ** **** video:

 

************

**** ** ********* ** any ****** ******* ******** 6.50 ** *****, ***** covers *** ******** ** Axis' ******* ******* **** (60%+). *** ****** *** shipped **** **** ********* from *** *******, ***** it *** ** ********* via *** *** ********* or**** ****** ********* ******.

*** ***** ***** **** not **** ******* ** this ********,**** *******.**** ***** *********, ****** **** reduced *********** ******** ** VMD4.

Configuration ********

************* ** **** ** ******* to ****, **** *** same ******* *** *******/******* zones, *****-***** *******, ***** objects, *** ******* *******.

** ****** ***** ************* options ** *** ***** below:

Few ***** ****** *******

*******, **** *** *** suffer **** ***** ****** caused ** ******* ***** affect **** ***** ******'* VMD *** *********.

*** *******, ********* ****** and ******* ***** ******* (e.g., ****** ********* ***/** in * ****) *** not ******* *** ****** in *** *****, ******* common ******* ** **** analytic *****.

***** ******* **** ******** monitors *** ****** *****, common ***** ***** ******* in **** *****, *** not ******* ****, ***** below. 

Outside ***********

********, **** *** *** alarm ** ******* **** as ******* ********/******* ***** commonly ******* ***** ********* we **** ******.

*******, ***** ******* ** the ********** ** *** scene ********** ********* ****** (shown *****). *****, *********, and ******* *** ****** multiple ******. ***** ***** be ******* ******** ** adjusting ***** ****** ********, but *** **********.

**** **** ***** ****** settings ****** ** ********* set ** ****** ******** detection ** *******, ** distant ***** ******** *** be ******* **** ***** or ***** ******* ** the **********, *** ******* by *** ********, ******* missed ******.

******, **** *** **** caused ******* ***** ******, as ****. **** *** most ******, ******* ******** alerts. ****** **** **** could ** ******* ** increasing ******* ***** ****, but ***** ******/******* ****** settings *** ** ******.

**** *** **** ******, but ******** ** *** cameras' ***** ********* ****** alarms, ********** ** ********. 

*******, ******* ****** ****** in *** **** ********* alarms ************, ****** ** a **** ***** **** than *******/*******.

 

Camera *********** ***********

** ****** **** ***** several ******, ********* ***-*** * series ******, * ******, and * ****** *****. In *** *****, *** cameras ******** ******** ******** at **-** ***, **** the ***** ******* ********* on ********** *** *** width (*** ********* ***** later). 

*******, *** ******* ****** included ********** **. ***** without ********** ** *** ****** to ******* ***** ** nighttime ****** ******* ******** as ****** *** ****** too ****. 

Detection *****

**** ************ ******** ******** at ~**-** *** **** day *** *****. **** results ** *** ********* outdoor ******/*** ****** ** cameras ******:

  • *****-*** (*** **:*): ~***' **** *****/~***' ********
  • *****-*** (*** **:*):~***' **** *****/~**' ********

 

Improved *********** ******** ** ****

**** ********** ******** **** many ** *** ***** VMSes, ********* *****, *******, and *********, ** *********** over ****, ***** ********** only ** ******* *** ****' *** ****** ******* ***

VMS ***********

****'* *********** **** *****, Genetec, *** ********* ******* specific ***** ** ****** recording/alarms *** **** ******, varying ** **********, ********* below. ******* ***** *** released ******* *** **** in **** *.*, *** yet *********. ******** **** *** support **** (******, ** course, ******** ******** *** own *******).

Genetec/Milestone ***********

******* *** ********* *********** require ***** ** ****** VMD4 ** *** ******'* motion ****** **** *********** camera **** ******, ******* of ******** ***:

Exacq *********** *************

*** *** ******* **** VMD4 *** ** ***** legacy ****** ***** ********** (1, *.*, *), ***** recognizes **** ** *** default ****** ********* ******.

*** ******* ***** **** previously ********** **** ******* version ** ****** *********, users **** ****** *** camera **** *****, ****** the ***** ****** ***** using ***** ******, **** create * **** **** and **-*** ** *****. Exacq ********* ******* *** VMD4 ****. 

Other *****

*** ***** ***** ***** *****, it *** ** ******** to ******* ********* ** VMD4 ****** ** ********** the ******'* ****** ** its ***** (** *********) and ********** ********* ** input **********. ** *** also ** ******** ** trigger ****** ***** ******* TCP/HTTP *******. ** ******* both ** ***** ******* in*** **** ******.

Versions **** 

*** ********* ******** ******** were **** ** **** test:

  • **** *****-*** ** **: 8.10.1
  • **** *****-***: *.**.*.*

Comments (25)

Since those trees are bare now, do you plan on revisiting this test when there are leaves on the trees? I am wondering if the leaves will impact the accuracy.

We tested Axis and Hikvision together over the past few weeks. Hikvision will be released next. As we're testing others later on, we'll likely do some retesting when leaves are on the trees.

look forward to seeing this :)

I bet this works with Avigilon when you configure Arbitrary Events in ACC. 

I can confirm VMD 4 works with Avigilon ACC 6.8.4.0 using Arbitrary Events.  You can use VMD 4 to trigger any alarm or rule in ACC and trigger recording.  Tested using Axis P3228-LVE and Q1615-MkII both running firmware 8.10.1

 

 

Sorry, this took so long.  Here is an image of searching VMD4 events in ACC.  Nevermind the "Axis VMD3 test" as we were testing with VMD3 and VMD4 but forgot to change the name. 

 

Is the variable object size detection global to the FoV? Or can ROIs be established that have different thesholds? If yes, then you can overcome some of the challenges of objects getting smaller as the camera to object distance increases.

Object size is unique to the profile, which essentially amounts to an ROI. You can only have one detection zone in a profile (but more than one exclude zone). Settings for short lived, small objects, and swaying are configurable independently of other profiles.

Great ,  detailed post!  For all the related trash Google results out there , this site is the only place you'll find sold,  proper reviews/ comparisons like this. 

 Two other important notes/upgrades with axis VMD4 as compared to vmd3 ;

With 4, you can finally use chrome (and I assume Firefox),  to configure your motion Squares/polygons, in VMD 3 you had to use Internet Explorer plus axis s ActiveX component  (I think it's called access media control or something similar).

Additionally an idea:  during your IPVM video screen grabs, you can add an event (on the cam) with action=overlay text , " Motion detected" , trigger= Any motion detection profile.   This will show when md is triggered. Not a big deal, but this event is something We like to add to deployments as it's easy to see on remove streams how often motion detection is being tripped (and it's accompanying high-quality low FPS short record to SD card) 

I Should have added- if u use action=overlay text ,  The variable you should use on your text overlay settings is #D

(boy did this have me stumped for months..  I was sure it was a bug with the action , as #D isn't very clearly documented)

Very nice...I was also stumped on this one, thanks Joey. -mark

Generally love this.  But on the 1st video, after the lights are out you can see a 'perpetrator' in the doorway.  Was that alerted or no?  So if someone walked right there, clearly a person, would that have been caught?

The great trick is missing that person when conditions aren't great....when it's raining, dark with lights out, etc..  We've tested 3 good ones (not all from camera manufacturers) and they've missed a guy but been absolutely great 90% of the time.

It's not easy and Axis has a great version going.

It didn't alarm on that reflection of the person in the doorway, nor would I think it should, actually. If someone just happens to walk by in the hallway, it's not actual motion in the room, and shouldn't alarm.

As far as detecting in the room, yes, VMD4 triggered very reliably in that scene and all the interior scenes we tested. We will grab a clip of that.

That's the great trick.  That -is- an actual person, and if that's my warehouse or clubhouse and someone stopped at the door or looked in I'd want to know.  It is a real live person.  I agree 100% that it generally works great inside, and we find very good outside as well.  Like anything else it falls up to the integrator to have it all set right so not to miss that critical event.

Great review

Are there any motion detection analytics out there in which you can create multiple detection areas with unique min/max pixel size within the same camera profile so you can catch humans far away but not small animals close up?

Thank you so much.  We just got bunch of new Axis cameras to install.  Axis manual is very weak.  It's online help in the camera gui does not go in details as you do.  Their webinar on Zipstream isn't very good either (refers to web gui help!).  Genetec manual on configuring Axis zipstream products is weak too.  

This is the best how to I've found so far. Thank you again.

 

There is one error I am getting that perhaps you can help? When I go in to Axis Video Motion Detection 4, error message below comes up on the top right-hand corner.

"Other Axis video analytic applications appear to be running already.  For performance reasons, it is recommended to run only one at a time.  Please stop: AXIS Video Motion Detection 1."

Under "Apps" there is only one app that is installed - AVMD 4.

 

 

Seeing as this is a new axis cam , my first guess would be it's your nvr trying to run edge motion detection incorrectly (or other onvif related mess up)?  

See if fully removing the cam from your NVR changes anything.  Otherwise look into axis plain config menu (for random settings / motion windows(

(or as a last resort , factory reset cam, wo nvr attached)

Thank you Joey.

 

 

Thanks for this report. I hope to see even more tests and reports on Video Analytics here on IPVM. This is a field where its a big gap in the claimed performance/the expectations of the clients and the actual performance of the products.

I wonder, is this really a test of Axis onboard analytics, or just the latest Video Motion Detect from Axis?

Looking at their website it seams like they have analytics packages that could perform a lot better than VMD4:

https://www.axis.com/products/analytics-and-other-applications
https://www.axis.com/products/axis-loitering-guard
https://www.axis.com/products/axis-fence-guard
https://www.axis.com/products/axis-motion-guard

I also really like to see a test of video analytics from VCA Technology.

http://www.vcatechnology.com/

I wonder, is this really a test of Axis onboard analytics, or just the latest Video Motion Detect from Axis?

Axis calls VMD4 a 'video analytics application'. However, to your point, there are more advanced options. We do plan to test them (e.g., fence guard) but we are starting from the entry-level, free ones to first understand how well or not they work. We'll be doing more video analytics testing over the next few months.

As for VCA, we might test them but that is a lower priority relative to larger companies like (alphabetically) Avigilon, Axis, Dahua, Hanwha, Hikvision, etc.

Good material here on VMD3 vs. VMD4.  Does anyone know what the differences are between motion guard and VMD4?  They look pretty similar...Also, note, on a Q3517, I have fence guard, motion guard, VMD4, installed. 

I realized digital auto-tracking doesn't work with this cam - not sure why?

-mark

I actually just asked Axis for clarity on performance differences between VMD4 and Motion Guard, as we are planning an updated test of that soon. I'll post back when we get a response.

As to Digital Auto Tracking, we asked in the past about some missing models, because several of the high end/high resolution cameras are missing from the compatibility list. According to Axis, the app was unstable on some models, so isn't listed as supported, but this was expected to be improved in later firmwares. I'll see what they say on this topic, too.

Thanks Ethan.

Update: We tested Axis' latest VMD4, 4.4.2, released in Axis firmware version 9.60. This test was originally conducted using VMD4 v4.2.

Performance is generally similar. We saw no issues on blowing foliage, similar to our original test, slightly stronger against shadows cast by the trees.

However, it still had issues with animals, such as the large dog shown here:

And it suffered from false alerts caused by headlights outside the field of view, shown here reflecting on the side of the trailer.

As Axis releases new versions we plan to retest periodically and will update here.

Read this IPVM report for free.

This article is part of IPVM's 6,584 reports, 886 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now

Related Reports

Avigilon Social Distancing Analytics Tested on Aug 26, 2020
Avigilon released its social distancing analytics in response to the...
Uniview Deep Learning Camera Tested on Jul 14, 2020
Uniview's intrusion analytics have performed poorly in our shootouts. Now,...
Avigilon UMD / UAD Tested on Oct 14, 2020
Avigilon's Unusual Activity Detection and Unusual Motion Detection claim to...
Hanwha AI Analytics Camera Tested on Aug 11, 2020
Hanwha has released their Wisenet P AI camera, adding person and vehicle...
Hanwha Face Mask Detection Tested on Jul 01, 2020
Face mask detection or, more specifically lack-of-face-mask detection, is an...
Single Frame Gait Recognition From Michigan State and Osaka University Examined on Oct 01, 2020
Gait recognition has the potential for accurate identification at a distance,...
Verkada 2020 Cameras Image Quality Test on Oct 06, 2020
Verkada's first-generation cameras suffered from numerous video quality...
Avigilon Elevated Temperature Detection Camera Tested on Sep 17, 2020
Avigilon has entered the temperature screening market with the release of...
Avigilon Face Mask Detection Tested on Jun 24, 2020
Face mask detection or, more specifically not wearing a face mask, is an...
Clinton Public View Monitor (PVM) Mask Detection Tested on Jul 09, 2020
Face mask detection, or more specifically not wearing one, is expanding...
Axis Compares Fever Camera Sellers to 9/11 on Sep 18, 2020
Axis Communications, the West's largest surveillance camera manufacturer, has...
Verkada Video Analytics Tested 2020 on Oct 08, 2020
In 2019, Verkada released people and face analytics but our testing then...
US Startup Fever Inspect Examined on Aug 03, 2020
Undoubtedly late to fever cameras, this US company, Fever Inspect, led by a...
FLIR Screen-EST Screening Software Tested on Jun 30, 2020
In our FLIR A Series Test, the cameras' biggest drawback was their lack of...
China Bems Temperature Measurement Terminal Tested on Sep 22, 2020
Guangzhou Bems (brand Benshi) is the manufacturer behind temperature...

Recent Reports

Panasonic Presents i-PRO Cameras and Video Analytics on Oct 19, 2020
Panasonic presented its i-PRO X-Series cameras and AI video analytics at the...
Augmented Reality (AR) Cameras From Hikvision and Dahua Examined on Oct 19, 2020
Hikvision, Dahua, and other China companies are marketing augmented reality...
18 TB Video Surveillance Drives (WD and Seagate) on Oct 19, 2020
Both Seagate and Western Digital recently announced 18TB hard drives...
Watrix Gait Recognition Profile on Oct 16, 2020
Watrix is the world's only gait recognition surveillance provider IPVM has...
Intel Presents Edge-to-Cloud Ecosystem for Video Analytics on Oct 16, 2020
Intel presented its processors and software toolkit for computer vision at...
Best Manufacturer Technical Support 2020 on Oct 16, 2020
5 manufacturers stood out as providing the best technical support to ~200...
Microsoft Azure Presents Live Video Analytics on Oct 15, 2020
Microsoft Azure presented its Live Video Analytics offering at the September...
Worst Manufacturer Technical Support 2020 on Oct 15, 2020
4 manufacturers stood out as providing the worst technical support to ~200...
Clorox Announces, Then Pulls, Fever Camera on Oct 15, 2020
For almost one week, Clorox was marketing fever cameras. The booming...
Faulty Hikvision Fever Cam Setup at Mexico City Basilica and Cathedral on Oct 14, 2020
Donated Hikvision fever cameras (claiming screening of 1,800 people/min. with...
Directory of 209 "Fever" Camera Suppliers on Oct 14, 2020
This directory provides a list of "Fever" scanning thermal camera providers...
Avigilon UMD / UAD Tested on Oct 14, 2020
Avigilon's Unusual Activity Detection and Unusual Motion Detection claim to...
Longse Promoting Hikvision Partner Fullhan Chip Based Cameras on Oct 14, 2020
With Huawei HiSilicon production being shut down at TSMC, camera...
Meridian & Goodview (BEMS Relabeller) Temperature Screening Tested on Oct 13, 2020
A lot of temperature tablets look exactly alike and that is because they use...
Monitoring Alarm Systems From Home - Innovation or Danger? on Oct 13, 2020
Remote monitoring by alarm companies since COVID-19 is bringing cost savings...