Hikvision Video Analytics Tested 2016

By Ethan Ace, Published Feb 04, 2016, 12:00am EST

We tested Hikvision's video analytics on six Hikvision models, ranging from entry level to high end Hikvision cameras, in this outdoor test scene:

We tested against Avigilon's advanced analytics and Axis' Video Motion Detection 3 to see how they compare.

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

*********'* ********* *********** *** false ***** ********* **** solid **** ********** *** the *****, ****** **** using ******* ********. *********** of *** ***** *** high *** ******* *** similar ** ** ****** than **** *** *, though ****** ********'* ********** analytic *******.

*** *** ******* *** TVI **** ******* *** ******** coverage ****** ** ******* tested, ****** ******** *** false ***** ***** ******* to ****** *** ******.

***** ********* *********'* ***** analytics ****** ** **** to ****** **** ********* and ****** ** ******** properly *** ***** *****, as ******* ** ** so **** ****** **** to ******** ******** ******. ***** should ** ******* **** adjusting *********** ** ****, as **** ****** *********** may ************* ******** ***** alerts ** ******** *****. 

***********

*** ********* ** ***** on-board ********* ** ** charge ** **** *** cost ****** (****** ******* to *********/*********) ***** *********'* positioning ** **** **** simply * ***** ****** brand. **** ****** **** in **** **** *****, Sony, *******, *** ****** who **** ******* ***** analytics ** ***** ******* for **** ****, *** puts **** ***** ** Axis, *** ****** **** VMD */*.* (******** ******/*********) and ********* ********* (****** at ** ********** ~$** fee *** ******).

**** ****** **** **** Hikvision ** **** ********** where ***** ********* *** included ** * ******** item.

Key ********

**** *** *** *** findings **** **** **** of *********'* ** ***** analytics:

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

  • ** *** *******, *********'* analytics **** ***** ** numerous ***** *********** *** to ****** ******** *** vegetation, *** *** ******* or ****** ****** (**********, flashlights, *******, ***.)
  • ********* ********** **** *** region ** ******** ********** nearly *** ***** *********** from ***** *******.
  • ** ***** *********** **** triggered ** *** **** and **** ******* *********, similar ** ********'* *** analytic ******, ***** **** *** 3 ********* ***** *****.
  • **** *** **** ************ than *********** ********* *** to **** **** **** inclusion ***** ******* ** polygons, *** *** **** of ******* *****.

********* ********

********* ******** ****** ** camera. **** ***** ******** coverage *** ********* *** competitors:

** *********'* ******* ******, the **** ********** ** dome *** *********** */*" super *** ***** ***** performed ****, **** *** largest ******** ****, **** the **** *** ** the *** ********** ******* cameras.

*******, *** **** ************ all ***** ******* ****** at *****, ******** ********* subjects ****** * ***' HFOV/185' *****, ****** ***'/***' in ********, *** **** best ******.

***********

  • ******** ****** ********** **** Genetec ******** ******, *** not ******** ******* ******, ExacqVision, ** ********* ********. No ******** ***** *** embedded ** ***** ***** third-party *****.
  • ****** ********* ******** ** Hikvision's ****-**** ******, **** ROI *** ******** ***** overlaid ** ***** *******.

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

*** ******** ** *********'* current ****** *** ******** line ******** ********* ********* and ******** ********* ** no ****. ****** *** series ******* ********** *********, as ****, ***** ** this *****:

 

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

***** *** *** *********** to ************* ***** **** detection **** ***** **** difficult **** ** **** competitive ******:

  • **** ***** ******* ****:******* ******* *** **** only **** *****/******, ****** odd-shaped ******* ********* ** draw ** ********* ******** regions.
  • ** ********* *****:*** ********* ** **** not ******* ********* *****, only *******, ********* ***** to *** ******** ********* zones ** ***** * full ****** *** **** objects **** **** ** exclude.

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

***** ** *** *****, we ********* *** ********* settings ** ****:

  • *********** **:******* ******* ** * sensitivity ** ** ** all ******** *****, *** we ***** **** ** be *** **** *********** of ***** *** ***** alarm *********. ******** *********** quickly ******* *****, ***** raising ** ********* ***** activations.
  • ********* >* ***:** *** **** ******** of ************, ******* **** be ******* ** *** region ** ******** *** greater **** *-* *******. Raising ********* ** * or ****** ********** **** false *********** (********* *****-***** due ** ******* *********) while *** ********* ********. Threshold ******** ** *, meaning ***** *** ********* instantaneously.

** ****** ***** ****** and **** ** **** video:

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

***** ******* ********, *********'* analytics ******** **** **** amounts ** ***** *********** from *** *******:

********/*******

*** ******* ****** ** nuisance ****** ** *** was ****** ********/*******. **** even * ****** ******, branches triggered ********* ********* ***** near **********. **** ** these ****** ***** ** filtered ** ********** ********* to * ** ******, though *** **** *******, these ***** ****** ** removed **** *** ****** of ********.

*******

****** *** ***, ******* cast ** ********, ******** wires, ** ****** ******* in ***** ***** ****** some ***** ******, ****** at * **** ***** rate **** **********.

Detection ********

******** ******** **** ****** depending ** *** ********* camera *****. *** **** series (**** *** **** in *** ****) ********* best ** ******* ******, with ******** ***** ******* to ********'* *** ******** camera, ***** * ***' HFOV. *** *********** ****'* daytime ******** *** *******, ~240' ****, ******* ** the **** ***** **** VMD *.

*******, *** ****'* ********* ***** was *** **** ** all ******* ******, ******** a ~***' ********** ***, only ********** ******* **** its ******* *****. ***** cameras *********** *********** ***** in ******** ** *** light.

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

******** ***** ** *** 2000 ***** ****** *** TVI ******* **** **** smaller ******** ** *** 4000 *** **** ******. During *** ***,

** *****, ******** ******* significantly,

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

** *** *****, *********'* analytics ********** **** ******* Security ******, *** *** Milestone ********, ******** ******* Center, ** *********** ************. Bounding ***** ****** ******* are *** ******** ** video ** *** ** these *****.

*********** **** ******* ******** no *************. ****** *** simply ****** **** *** camera ** *** ***, and ******* ** *** event ****.  

****-****

*********'* ********* ********* **** their *** ****-**** ***, as ****. ***** **** is *********, ** **** as ******* ** ******** and ****** ******** *****, not ***** ** ***** party *****. 

Comments (7)

Thanks, Ethan, for another great, detail-oriented article. I wonder if defocus or scene change analytics might apply to flashlight blinding incidents?

Agree
Disagree
Informative
Unhelpful
Funny

Thank you Ethan for testing and most of all thank you Hikvision for once again supplying the video industry with IP Video cameras and VMS LICENSE FREE.

Agree
Disagree: 1
Informative
Unhelpful
Funny

...VMS LICENSE FREE.

OLD VERSION.

Agree: 1
Disagree: 1
Informative
Unhelpful
Funny

The discussion included iVMS 4200. Version 2.4.3.1 is not old.

iVMS 5200 is a completely different platform

Agree
Disagree
Informative
Unhelpful
Funny

iVMS 5200 is a completely different platform.

Sure they are different, 4200 is free and 5200 is far better.

As for versions, iVms 4200 is at v2.4.3.1, yet iVms 5200 is already at v3.2.010

Why? Maybe because they started 5200 at v3.0.

Agree
Disagree
Informative
Unhelpful
Funny

do you know if face detection is working on people covering their head with ski mask or burka? I mean which algorithm is used for face detection, like the Bosch IVA, which recognize the shoulders and head silhouette or does it detect by eyes nose and mouth?

Agree
Disagree
Informative
Unhelpful
Funny

I would be very suspicious of a vendor claiming to do accurate FR on subjects covered in a ski mask or burka. This problem is hard enough to solve with complete subject cooperation, at least where meaningful accuracies are required.

Agree: 2
Disagree
Informative
Unhelpful
Funny: 1
Read this IPVM report for free.

This article is part of IPVM's 7,324 reports and 971 tests and is only available to subscribers. To get a one-time preview of our work, enter your work email to access the full article.

Already a subscriber? Login here | Join now
Loading Related Reports