Avigilon UMD / UAD Tested

By Rob Kilpatrick, Published Oct 14, 2020, 10:05am EDT

Avigilon's Unusual Activity Detection and Unusual Motion Detection claim to "uncover events that may go unnoticed", but how does this claim stand up in real world scenes?

IPVM Image

We tested Unusual Activity Detection and Unusual Motion Detection for over a month, examining:

  • Do they both alert on proper unusual activity or motion?
  • Do they miss unusual activity in the area?
  • Do either UMD or UAD false alert?
  • How easy are UMD and UAD to setup?
  • Can they be trained to not identify usual activity if it considers the activity unusual?

*******

***** ** *** *******, Unusual ******** ********* (***** uses********'* ** *********) ********** ******* ** atypical ******** ** *** scenes ******, ******* *** out, **** **** ********** false ****** *** ** parts ** ****** ***** detected ** **** ** due ** ********* ** parked ******** ** *** tests.

*******, ******* ****** ********* (which **** ***** *** data *** ** *** most ******* ******** ******* perform ****) ******** **** significant ***** ****** ****** by *******, **********, *** lights ***** ****** ** and *** ** *** testing, ****** ****** ** fundamental ************ ** ***** motion *********, ***** ** is ***** **.

*******, ******* *** *** UMD ****** ***** ** simply **** ******** ****** as *** *******, ** activities **** *** ******* but ********** (**** ** emptying ***** **** * dumpster ******) **** ******* UAD/UMD ******.

Avigilon ********

******** ******* **** ******** on ***/*** ***********:

***** *** *********, *** and ***, *** ******** to ******* **** ******** events. ***** *** *** events **** ** ******** should ****** ** ***** for *** ***** *** meant ** “**** ******” or ************* *** **** are ********* ** *** Focus ** ********* ********* (FoA). ** *** **** there ****** ***** ** a ****** ** * secure **** **** *** up ** ‘****** ** Area’ **** – ** is ********** *** ******** and ***** **** ** very *** ***** *********.

******** ******** *** ***** of ********* ********* ** the ******* ***** ** deliver ******* ******. ***** events *** ************* *** therefore * ****/***** ** not ******* ** *******. As ***** ***** ***** events **** ** ** more ******** *** *** meant ** ** ******** quickly ** ** ******** in *** *** *********. Users *** *** ***** judge – *** ** providing **** **** ********* (and *** ******* *** them ** ******* **** cameras).

Unusual ******** ********* **. ******* ****** *********

*** *** *** **** similar ****, *** *** uses *** **** ******** analytic **** ***** **********'* ** *********(*.*., ***, *****, ****, classifications) *** *********** **** is *******, ***** *** uses ***** ****** ********* events, ***** *** ** nature **** ********. ** our *******, ***** *****, UAD *** ************* **** accurate **** ***.

******* ******** ********* ** available **** ***** ******** *******. *** ** ********* on****,***,***,****, ***** **** ****.

Accurate ******* ******** *********

******* ******** ********* ********** triggered ** *** *****. In *** ******* *****, it ********** *** ***** as ******* (**** * yellow ******** ***) ** they *** ****** ****** in **** ****.

IPVM Image

*** **** ********* ** unusual *****. ** ** area ***** ***** ******** normally ****, ** ***** the ******* ******* ***** as *******.

IPVM Image

*******, ****** ******** ** unusual ***** ** *** scene **** ******* **. In *** ***** ***** below, ******** ********** **** from *** ****, ******, to *** ******* *** to *** *****, *** rarely ***** *** **** near *** ****** ****.

IPVM Image

People ******** *** ********** ******* ********

***** *******/**** ***** *** properly ********, ****** ******** at **** ***** ******* the ***** **** *** identified ** *******.

IPVM Image

Vehicle ********* ******** ***** ************** ***** ***** *** ******

** **** *********, ***** of * ***** **** considered * *** ** an *** ***** *** were ******* ** ** unusual ********.

IPVM Image

************, **** ****** *** considered * ****** *** created ** ******* ******** event. **** ******** **** once ****** ******* ***** of *******.

IPVM Image

Normal ******** ********* ********** *******

****** ******** **** ** a ***** ******* ******* the *** ****** *** day *** ******* **** sometimes ********** ******* ****** not ********** ******* *** that **** ** *** day ** **** ** the *****.

Unusual ****** ***** ***********

******** ****** ** ** area *** ********* ******* on **** ** * person ****** ** * low ******* **** ** the *****.

False *** ****** ** ******* *** *** **********

*******, ***** *** ****** on **** ******** ***** motion *********, ** ******** from ******** ***** ****** caused ** ******** *********** not **** ** ***.

*** *******, **** ********** highlighting ***** ** *** scene ***** ** *** typically *** ****** ********* UMD ******.

IPVM Image

** *** ******* ***** cast ** ***** *****:

IPVM Image

*******, ***** ***** ******* (e.g., ******* ****** ** and ***) ****** *** events.

Significant **** ****** ** ***** *****

**** ******* ******** *** Unusual ****** **** *********** periods ** **** ** learn *** ***** *** start ****** ******** ******. UAD ***** ***** * week, ***** *** *** require ** ** *** weeks. ****** *** ***** is ******** *******, ****** occurrences **** ** ******** walking ******* *** ***** below *** ******* ***** alerts:

******* ** ********'* ***** self-learning *********, ***/*** **** a ******** *** ***** shows ******* ******** ********:

IPVM Image

No ****** ******** ** *** ** ***

**** ******* ****** *** Unusual ******** ***** ******** in *** ***** **** a *** ****** ** time. ***** *** ***** is ******* *** **** will **** *** ****** of **** **** ********* deem *******, ***** ** no *** ** ***** the ********* ** *** trigger ****** ** **** it ********* ******* *** the **** **** ***, such ** ****** ******** trash **** *** ******** shown *****.

IPVM Image

Focus ** ********* *********

***/*** ****** *** ********** to ********'* ***** ** Attention *********, **** ****** highlighting *** ********** ******'* hex ***** ****. ********* can ****** ***** ** view *** **** ***** or ****** ***** **** to **** **** *****.

** ***** **** *********** unusual ****** *** ******** detection ****** ** **** visibly ********* **** ******* a ********* ***** **** with * *****-****** ****** because, ** ******* ** ******** **** testing, *** ****** *****/******** are *** ********* ** a *** **** ******* the ******* ********* ** a ******** ********. ********, other ******** ****** **** facial *********** *** *****-***** the **** ** ***/***, which ****** ** *******.

Simple ***** / ***

**** ****** ***** ** required *** ***/***. ** H5A *******, *** ** enabled *************. ** ******* supporting ***, ***** ****** toggle ** ** **** a ********.

****** *** ***** ** Avigilon's ***** ** ********* interface, *********** ** * lighter **** **** ******* events *** ********. ****** in ******** ***** *** be ***** *** ***** search, ******* ** ***** analytic ******, ** *** be ********* ** *** timeline *** ****** ******.

*******

******** *** *** *** are ******** **** ********* of ******** ******* ******, no ********** **** *** required.

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

*** ********* ******** **** used ****** *******.

  • ******** ******* ******: *.**.*.**

Comments (4)

Avatar

Ryan Ritter

Avigilon | 10/15/20 07:23pm

************, **** ****** *** considered * ****** *** created ** ******* ******** event.

***** ** *** *** Spider *** ********* ** working!

** *** ***********, * am ******* ***** *** truck ***** ***********. ** you ********* *** ****** in **** ******* *** get **** ***** ** parking *****/***** ****** *** end ** *** **** ramp ** **** ******?

****'* ** ******* ***** for **** *** **** triggered **, ********** ** the ***** ******* * couple ******* ***** ****. Reminds ** ** *** initial ******* **** ** showed ** *** **** a ****** ***** *** with ***: ********* ** a ****** (*****, ***, so **** ****'* **********) walking ** * **** and ******* ** *** out, ********* ** ******.

***** ***, *** ****** it ********* *** **** covered *** **** *** the ****** ********** *** areas ** *** ****. So ** ***** ***** and **** *********** ** the ****** ****** ******* the ****.

***** ****. ******* ******* through **.

Avatar

Ethan Ace

IPVM | In reply to Ryan Ritter | 10/15/20 08:17pm

** ******** **** ****** drive ****** **** **** exact ***** ***** ***. The ******* **** **** does ************ ** ***** is ***** ******* ********** from **** ********* ** the ****** ** *** night. **'* ******** ** could ** ****** ** the **** ** ********, though, *** ** ******* pretty *******. *'* **** wondering ** ** ***** become **** "******" ***** more ****, ** *** camera ******* **** ******, etc.

Avatar

Ryan Ritter

Avigilon | In reply to Ethan Ace | 10/15/20 08:25pm

* ***** ******* *** answer ***** ********** ** yes, *** **** *'* unsure ** ** *** many ********* ** ***** take *** *** ***** to ******** ** ******.

**** ********/******** ** ***/*** is **** ** ** wishlist *** ***** ********. I ***** ****'* ****** what **** ***** ** to ** ****-*****.

****** *** *** ********!

Michael Miller

In reply to Ryan Ritter | 10/15/20 08:43pm

****, **** ** * message * ***** ** happy ** **** ***** UMD *** ***.

Read this IPVM report for free.

This article is part of IPVM's 6,651 reports, 895 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
Loading Related Reports