Hikvision DeepinView Camera Analytics Tested

Avatar
Rob Kilpatrick
Published May 08, 2019 12:58 PM

********* ** ********* *** '**** ********' offerings **** * *** ****** ****** called '**********' ******** ***** ***** ********* *** improved ********.

IPVM Image

** ****** *********** ***** ****** ****** (**-*********-****)*** ****** ** ******** ****** *** over *** ***** ** *** *** it ******** *********:

  • ***** ***** ***********: **** *** *** did *** ***** ***** ******?
  • ********* ********* ********: *** ** **** valid ****** ** ******** ** *** scene ** ******** **********?
  • **** ********* ***********: *** **** ********* work ******** ** **** ***** *** low ***** ** ******** ******?
  • **** ** *****: *** ********* ** time-consuming *** ***** *** ************?
  • **. *********** *********: *** *** ********** compare ** ****** ********* **** ***** ****** ********* ******** - ********, Axis, *****, *****, ******, *********, *******?
  • **** ********** **** ** ***? *********, Huawei, *****, ******, ******?
  • **** ********** **** ******** ******* ****?

Performance *******

***** ** *** ***** ** *******, reviewing ********* ** ******, *** ********** DS-2CD7A26G0-IZHS *** ******* *********:

  • ** ****** *********:****** ******* *** ****** ******* ** avoid ***** ****** ** ******* ** the ***** **** ******* ********, *****, etc., ****** ** **** ***** ******** cameras.
  • ** ****** ** ****: *** ********** ****** *** *** alert ** ***** **** ** *** scene, *** ** *** **** ****** sources ** ***** ****** ** *** past ********* *****.
  • ** ****** **********:****** *******/******* ******* *** ***** **** not ****** ****** *** ***** ** testing ** ******* **********, ***, *****, and ** ****.
  • **** ********* **** ******** ** **** light:** **** *** ******, **** ********* did *** **** *** *****, **** at ***** **° ******* ** **** partially ********.
  • ****** *************:************* ** *** ******** ******** **** enabling ********* ********* ** **** *********, with *** ****** ********** ** * full ****** ****.

*******, ** ******** **** ***** ****** on ******* ******* *** ******* **** detection *********** ** *** *****:

  • ***** ******* ******* ******** ******:** *** *****, *** ********** ****** regularly ******* ** ***** ******* *****, sheep, *** ****, * *** ***** per ***. ******* ******* **** ** birds, ****, ********* **** ******* ** but **** **********. ** ** ******* if *** ****** *** ***** ******* as ***** ** *******, ** ** does *** ******** *******.
  • ******* ***** *******: ******* ***** ******* **** ** lights ** * **** ******* ** or **** ********* ******** ****** ***** alarms.
  • **** ********* *********** ***** ** ****/*** light:** *** *****/**** ****** **** ** on, **** ********* *********** ****** ************, sometimes ********** ******* ********, ********** **** at ******* ****** ** ******* ****** the *****.
  • ******** ******* *********: *** ********** ****** **** ** testing (****** ******* ** ************) ******* with ******** *.*.*, ***** ********* ***** poorly, **** ***** ****** ** *****, shadows, ****, ***. *** **** ** later ********. ********* ******** ** *.*.**, available ** ***** ******* *** *** yet ** *** *** ****, ****** they *** ** ** ********* ** the ******* ** ***** ********* *** added. *** **** ******* *** ***** on *.*.** *** ** ********* *** users **** ******* ******** ** ******* to ****.

Vs. ******** ****-******** *********

********** *** ********* ********* ** ***** alert ******* ***** ********'* ** ****-******** analytic ******* ********. ** *** *****, an ******** ****** ** *** **** scenes *** *** ********* ** *** animals ****** ******* **** ** *******, nor ***** *******, *******, *******, ***.

Vs. ***** ********* *********/***

********** *********** ** *** ***** ** beyond **** ** **** *********** ********* ** **** ******, ********* **** (********* ******** *** Motion *****), ****** ******* *, *****, and *********'* *** ******** ********* ********* analytics, **** **** ****** ********* ** foliage, *******, ****, *** **** ***** animals.

***** **** ********, *** **** ******** from *** ******** ******** *********** ****** DeepinView ** *****'* ***, ***** ******** animals, ****, *******, *** ****, ****** occasionally *** ****** **** *******. ***** setup ** **** ******* **** *********, however, ********* ******* ***** ** ********* and ******* *********** *** **** ******, instead ** ****** ******** *** ****.

******** ** ******** ***, *********** ** drastically ******, ** **** ****** **** VMD ***** ****** ***** ** *** pixel ******, ********* *** *** ******* to *** ****** ***** ***** ******* above, ******* ************** ******* ******, ********, or ***** *******. *******, ********** ** far **** ********* **** ************ *******.

*******

*** **-*********-**** ***** *** ~$*** *** online. **** ******* ** ********** *******'* ********* **** **(******* ***** *** *****/******** ******), ****** moderately ****** ************'* ******** ******* (~$***-*** ********* ****** *****).

*** ***-** * ****** ********** ** this ****** (**-*********-****) ***** *** ~$*** ******, ****** deep ******** * ~$*** *** **.

Physical ********

*** ********** ****** ** ********** ******* to *********'* ***-**** ******** ****** ******. It ** ******* ** **** *** construction ** ***** **** **** *******, such ** ****** (***** *****), ******** longer, **** ** ******** **** ***.

IPVM Image

** ****** *** ******** ** *** camera ** **** *****:

DeepinView ********

*********** *** ********** ****** ** ******. Users *********** **** **** ** ****** Intrusion ********* (** **** *********). ***** settings *** ** **** ** ******* (full ****** ********* ****, *** *******/******* object **** ********), *** *** * limited ****** ** *********** ** *** tests.

** ****** *** ***** ***** ** this *****:

Rejected ***** ******: ****, *****, *******

*** ********** ****** ******** **** ***** alert ******* **** ******** *********** ** past ********* *****.

*** *******, ********** *** *** ***** on ***** **** ****** *** *******, through ******* ****** *** *** *****. This *** *** ** *** **** common ******* ** ****** ** **** tests, **** ****** ***** ******** ****** frequently ******** ** *********.

No-False-Alerts-In-Heavy-Rain-Wind

*** ****** **** ******** ****** ****** by ******* ********/******* ** **** ** the ******* **** ******, ******* ****** issue ** **** *****.

No False Alerts On Brush Shadows

Regular ***** ****** ** *******

*** ********** ****** *** ******* ****** with ***** ****** ** ***** *******, such ** **** ** *** ******** of *** ***** ** **** *****.

False-Alert-On-Dogs-In-FOV

** **** ******* ** ***** *** sheep (***** *****). ** ** ******* whether *** ****** ********** ***** ******* to ** ****** ** ********, ** it **** *** ******* ************* ** provide ****** ** ****** ********.

False-Alert-On-Sheep-In-FOV

Fewer ****** **** ***** *******

**** **** *** ****** **** ************ alerted ** ***** *******, ********* * few ***** ** *** *******, ****** at * **** ***** **** **** larger *******, ***** ********* ****** ***** time **** **** ** *** *****.

False-Alert-On-Birds-In-FOV

*******, ** ******* *** *** ** the ***** *****, ***** ***** ******** on **** ******* ****.

No-Alert-On-Cat-In-Scene

Alarms ** ***** *******

******* ****** ****** ** ***** ****** was ******* ***** *******. **** ******* in ******** ******, ********* ******** ****** a ************:

Alert After Lightning Flash

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

****** * **** ** ******* ** various ******, ****** ** **** **** not ****** **** ******** *** ***** of **** ** *** ****** ******* or *******.

No-Missed-Detections

****** ********* **** ********* **** ** heavy ****, ******** ** *** ****** walking ** *** ***** *** ******** the ****.

People Detection In Heavy Rain

Person/Vehicle ********* ***** ***********

** *** *****, *** ********** ****** was **** ** ******** ****** ****** at ***** ~** *** (~***' ******** using * ~***' ****). **** **** detected ** **** ** **** ********, ~530', ***** *-* ***.

**** ** ********** ****** **** ******** during *** *** (~* ***), ****** lower ** *****, ***** ******** ******** as **** ** ~** *** *** detection. ***** *** ******** *** *** night ** ~*-* *** ** *** tests.

IPVM Image

Face ********* ***********

*** ********** ****** **** ******** **** detection, ***** *** ** ******* ******* of ********* *********, *** *** *** cannot ** **** **************. **** ******* is ******* ** ********* ** ** in *** "*** ********" *** ** basic *************. "***** *****" ****** ** typical ******** ***** **** ** ********* detection *** **** ********.

Smart-Event-Or-Face-Capture,-Not-Both

** *** ******* **** ********* ****** well ** * **** ***** ***** (~200lx), **** ********* *** ******** **** at ******* ******* ** **** (~** degrees) *** ******* ******** ****** *** scene ******** **** **** ** *** face.

Face-Detection-Works-Well-In-Full-Light-Harsh-Angles

** *** *****, **** ** ** or ***, *********** *********, **** *** camera ******* ***** ** ******** ******* across *** *****.

Face-Detection-Misses-In-Low-Light-IR

Movidius ****** * *** ****

*** ********** ****** **** ******* ******** ****** * ****** ********** unit (***), ******** *****.

IPVM Image

******** ** ************* ***, ***** **** * ****** ***** Nvidia ********* *** ** ******** ** intrusion, ********** * ****** * ** a ****** ******* ** ***** ******** much ******* ********** *****. ** ******* move ** *** ***** ****** * in *** **** *** *****, ********** gains *** ******, ** *** ****** X ** ******* **** ******** **** the ****** * (*** *** ****).

Unknown ***

** **** **** * ******* ******* SoC, *** *** ********* ** ********* labeled ***** ********* **** ** ***** cameras:

IPVM Image

* ****** *********** ****** ********* * possible *******:

  • * *********-******** ***, ******* ** ***** the ***** ***** ***** **** **** historically **** ** ***** ******. "**** would **** **** *** ***** ******** they **** ** * ****** ** bargain ******** ******* ****** ********."
  • * ********* *** **** ******* *********, e.g., *********. ***** ***** ********** *****, Hikvision *** ** ***** ***** ** others **** ***** **** ** ****. "It **** *** **** **** **** volume ** ***** ***** **** **** your *** ****** ******** ** ****... so ** ***** ** **** * normal *** *** ***** ********* ****, but **** ********* ********. "

Versions ****

*.*.** ******** *** **** *** *** Hikvision **-*********-**** *** ** *** ***** of *** ********.

** **** ****** *.*.* *** *.*.** (what *** ******* ** ** *** is ***** ****** ** ********* *** as *******), ******** ** ***** ******** was ******* *****, ********** *.*.*, ***** alerted ** *******, *****, ****, *** more ******* ***** ***** ******** *** not. ** *** *** ******* *** examples **** *.*.* ** *.*.**, *** users ****** ** ***** ** *** substantial ********** ** ***********.

Country ** ******

*** **-*********-**** *** **** ** *****.

Comments (15)
U
Undisclosed #1
May 08, 2019

*** **** ** *****/***? ** *** the ****** *********** **** ** * range ** **** **** * ******** FOV, ** *** **** * ******** FOV?

Avatar
Ethan Ace
May 08, 2019

** *** *****, *** ********** ****** was **** ** ******** ****** ****** at ***** ~** *** (~***' ******** using * ~***' ****). **** **** detected ** **** ** **** ********, ~530', ***** *-* ***. ***** ******* are ***** * ~**° ***.

(2)
JH
John Honovich
May 08, 2019
IPVM

**** ** **** ********* ** *** test.

********* *** **** ******* *** *** helpful **** **** ***** (*** ******* them, **** ** * ****** ******* we ***, **** *** ********* *** trained ** *** / ******** ********). Tech ******* ******** **** *** ******** shown ** *** ** **** .** was *** ******.

**** ****, ********* ********** ******** ******** tech *******. ** ******* *** ** European ******** *** ********* **** .** was ******. ** *********'* ******, *** USA **** **** ** * ****** link ** *** .** ******** ** we ***** ****.

*** **** ****** **** ****'*********** ********** *** ****************** ******* *******.

******* ** ****, ** **** ******* firmware *** **********. *******, *** *** is **** ******* *** '*' ******** version ********* (***** ** **** ** bought) ***** ****** *** *** '*' hardware, ***** ** ********** ******. ** are ****** ** ******** ******* ****.

*.*. - ****** *********, ***'* **** until *** **** ***** ********* ********. If *** ***** **** ** *******, that ** **** **** ** ***, production ********* (** ***** ** *** USA) ****** **********-***** ********.

(1)
U
Undisclosed #2
May 08, 2019

******* *** ****** **** ******** ** common *** ******* *************. ** ** experience **** **** ***** *** ********* virtually *** ***** ******* ******** **** a **** ***** ** **** *** require * * *** ** *********** to *** ******** ** *** **** right. **** ** *** *********** **** use ***** **** ************ ** ****** integrators *** **** *** *** ********** of ****** *** ********** ***** ***** of ***** ******* ******** *****  ****** are **** * “****” ***** *******.

(2)
(2)
Avatar
Vincent Tong
May 09, 2019

*** .... ***** *** *** *** the ***** **** ?

(7)
UI
Undisclosed Integrator #3
May 09, 2019

*** ***** **** *** ***** **** into **** *********. *** ****** **** their **********? *** ** **** **** without *** *****, **** ** ****** knowledge? ****'* **** **** **** *** stored **** *** ** *** ** the ******* **** **** ** ** ethnic ******** *****? *'* *** ********* IPVM ** ****** ***** **** ********** responsibilities *********.

********* * **** ****** ****** *****. I'm ******* ****** - ****** *** two-week **** **** *** **** ******** took ***** *** *** *** ***** positives ******* **** **** - ****** the ********* *** ******* ***********?

*** ******** ***-*******, ***** ****** ****** sabre-rattling - *** ** ******* ****** the ***** ******* ******* ** *** camera? **** *******.....

(1)
(1)
JH
John Honovich
May 09, 2019
IPVM

****'* **** **** **** *** ****** data *** ** *** ** *** targets **** **** ** ** ****** minority *****?

* ***** *** *** ****** *** I ***'* ***** **'* ***** ** make *** ** ***** ****** ******. Accept **** ********* ** ********* *** it, **** *** *** ******* ************ ***** (**** * And *)

** *** *****, ******* ***** *** be **** ** ** ****** ************ ******* ***** *** ************ ***** ********* **** ***** ****** abuses *****.

UI
Undisclosed Integrator #3
May 09, 2019

* *** **** ** ***** *** you **** ** *** ***** **** your ***.....

************* *** ****'* ****** **** *** technical ***** ******, *** **** **** the *********** ** *** *** ******* out.

(2)
(3)
JH
John Honovich
May 09, 2019
IPVM

* *** *** ********** *** ****. That *** ***** **'* ***** ** joke ***** ***** ****** ****** **** a *** ***** ***.

*'** **** ***** ******* **** ******** about ***********. ** *** ******, ** you **** ** *** * ********* question, *** ** ******* ****** ***** oppressing ****** **********.

(3)
UI
Undisclosed Integrator #3
May 10, 2019

* ***’* **** ** ********* ** devalue *****’* **** ** ************ **** rant *** ********** **** *** ******.

(2)
JH
John Honovich
May 10, 2019
IPVM

*** **** ********** *** ******** ********* by ******** ** ***** ********* **** human ****** ****** ** * **** that *** ******* ** ** **** that.

(1)
Avatar
Ethan Ace
May 10, 2019

** *** ** *********** **** ****. It’s **********, *** *** ******* **** firmware ******* ** ******** *******, *** it *** *** **** ******** ** branches ***** ******* **** ** *** same ***** ** ***** ******** **** after ****** **** ** *** **** area. **** ***** ** ** *** to **** ****** ** ***** ** false, ** ** ***’* ** ******** trained. 

(2)
UI
Undisclosed Integrator #3
May 10, 2019

****** *****.

UM
Undisclosed Manufacturer #4
May 10, 2019

*** ******* *** ****** **  "****** Hisilicon **** ****" 

(2)
JH
John Honovich
May 10, 2019
IPVM

#*, ******. *** *** *** ****** that ***? ** ***** **** ******* or ******** **** ******* *** ** identify * ******** ********* ****?

(1)