Hikvision Acusense Analytics Tested
********* ***** "*** ***** ****** ** All" ** ***** *********** ******* '**** ******** **********' ****** ***** cameras *** *********. *** *** **** do ***** *** ****** ****?
** ****** *** ****** ************ **-*********-*/******** *****-*******-**/**/** ****** **** ***** *****, *********:
- ***** ***** ***********: **** *** *** did *** ***** ***** ******?
- ********* ********* ********: *** ** **** valid ****** ** ******** ** *** scene ** ******** **********?
- **** ** *****: *** ********* ** time-consuming *** ***** *** ************?
- **. *********** *********: *** *** ********** compare ** ** ********* **** ******* *********+******* ****** ****** ********* **** ***** ****** ********* ********?
- **** ********** **** ** ***? *********, Huawei, *****, ******, ******?
- *** **** **** ** ********* **** major *****?
******* ****** **** *** *** ***** of *********'*********** ** ********* ******************* ***.
*******
**** *** ****** ** ***** ***** of *******, *********'* ******** *** ******* positives:
- ** ***** ****** ** **** ******* Sources:******** ******* **** ** ******* ********/*******, shadows **** ********, *** **** ******** on *** **** ** *** ****** did *** ******* ***** ******.
- ** ***** ****** ** **** ****** Indoor *******:****** ****** ******* ** ***** ****** such ** ******* ***** *******, ******* moving ******* *** *****, *** ****** turning ** *** *** **** ******* throughout *******.
- ******** ********* ** ****** ******* ******* Or ********:******** *** *** **** ****** ******* upright ** ******** ******* *** ***** during *******, *** ** *****.
*******, ** ******* **** *********** ****** with ***************** *** ***** ******:
- ****** ********** ** ********:****** ******* ******* *** ***** **** sometimes ********** ** ********, **** ***** when ******** **** ** *** **********. This ***** ****** ** ****** ***** completely ****** ** ***** **** *** vehicle ****** *******, *** ********* ** detect ********.
- ***** ****** ** *******:***** *** ***** ******* (******* **** cats ******* ****) ******* ******* *** scene ********* ******** ***** ******, ********** of ******** ****. ******* **** **** often ********** ** ********.
- ***** ****** ** ****** ****:****** **** ** *** ***** **** sometimes **** ** ****** ******* **** passing ****** ****** ******* ** ********.
- ******** ***** ****** ** ******** ******* Windows:******** ******* * ***** ****** *** door ****** ***** ****** **** ** moving ******* ** *** *****.
- ****** *** ******** ******* ******* *****:****** ******* ******* *** ***** **** missed **********, *** *** *****, ******* and ***.
Hikvision ********
**** **** ********* *** ***** *** opportunity ** ****** ******* ****** ******* but *** *** ***** *** *********** for ******** ******* ** ******* *********. They ******* *** **** ******** ******* worked **** ** *********** **** ******** NVRs, *** ** *** ***** ** saw ** ********** ** *********** ******* Acusense *******' ** ***** ********* *** Acusense ******* **** ** *********** **** the ***.
Settings ******* ****** ** *** ******
** *** ******, ** ******** ******** to *** ** ********* *********** ********, but ***** **** ******* ******** ******** the **** *******. ** *********'* ********* analytics, ********* ********, *** **** ******** adjustments *** ********* (**** ****** ** in ***** ****** **********) *** *********** ("size ** *** ****** ***** *** trigger *****"). ******* ** ***** ******** eliminated ***** ******.
Vs. ***** *********+
*****'* *********+****** ************ ********* ********, **** ** false ****** ** ******* ** ****** which ********* *** ********* ********. ************* for **** ** *******, *********** **** requiring * **** ** ** *** and ****** ***/** ******* ********* ********.
Vs. **** **
****'* **** ** ****** ******************** ********, ******** *******, *******, ******, shadows, *** ***** *******. *******, **** has **** ***-******** *********** ***** **** it ********** *** ********** *** *** it ***** *********** ** ***** ***** VMS.
Vs. ******** *** ***** ***
********'* ****-******** ***************** ***(*** ******* ** ***** ****** ********* ********) **** ************ ********, ******** ****** from ******* *** ***** *** *** not ****** **** *** **** ************** issues ** ********.
Vs. ********* ******** ********* *** ***
******** ** *********'* ******** ********* *********, Acusense *********** ** ******* ******, ********* more ***** ****** **** **** *******, including ******* *** ******, **** ****** on *******, ***. *********** ** *********** better **** ******** ***, ***** ******* on *********** *** ***** ******, ********* but *** ******* ** *** ****** false ***** ******* *****, ******* ************** between ******, ********, ** ***** *******.
*******
*** ********* **-*********-*/** *** ** ***** online ***~$*** ******** *** **-*******-**/**/** (* *******/* ***/* channels *********) ***** ***~$*** *** ******.
People ******** ******* *******/********, *** *******
******** ******** ****** ******* ******* ********** testing **** ** ****** ****** *** detection ***** *** *** *****:
******** ******** **** **** ******** ** the ***** ***** ******* ********.
Running ****** ******
*******, ****** ******* ******* *** ***** were ****** **********. ** ******** ******* improved **** *********.
People ********** ************* ** ********
****** ******* ******* *** *****, ****** in *** **********/******* ****** *** ****** were ********** ********** ** ********. **** could **** ** ********** ****** ********** if ***** ****** ** ****** **** people.
False ****** ** *******
******** ********** ******* ** **** ***** and ***** ******* ** *** ***** of ****. ******* **** **** ***** classified ** ********, *** ********* ********** as ***** **** ******* ****** *** camera.
False ****** ** ****** ****
****** *******, ******** ******* ** ****** vehicles **** ***** ******** ** ***** light ******* ********, **** ****** *** vehicle *** *** ****. ***** *********+ also ******** **** **** *****, ****** we **** *** **** ** ** other *********.
False ****** ** ********
** *** *****, ******** ******* ** sunlight ********* ** *** ***** **** the ***** ****/*******. **** ******** *-* times *** *** ** *** ****** of *** ****.
No ***** ****** **** *******, *******, ** ************
********, ******** ******* ***** ****** **** common ******* **** ** ******* ******* and *** ******* **** *******, **** below.
******** **** ******* *** ** *** most ****** ******* ** ***** ******, heavy **** *** **** ******** **** the ****.
No ****** ***** ****** ** ****** ** *******
****** ******* ******** ******* *** ****** indoor ***** ****** **** ** ******* light ******* ** *** *****, ***** below.
******* **** ****** ****** ******* *** scene **** **** *******.
Detection ********
** *** *****, ******** ******** ******** at ~***' ********, ~** *** ** the *** **-*********-*/** ****** ** ******. These *** ************ *** ***, ********** to ***** ** ********, *** ***** than*********'* *** ********* *********, ***** ******** ****** ****** *** PPF *** ********* *** *** **** more ***** ***** ******.
Simple *****
***** ** *** **** **** **** Hikvision's ****** ********* ********* *****, **** requiring *** **** ** ****** ** people ***/** ******** ****** ** ******** in *** *****:
VMS ***********
******** ****** ****** ********* ***********'* ******* ********* *********** ***** ***** ***. ***** ****** were ********* ** ** ******* *** Milestone ********, *** *** ********* **** not ********* ** ***********, ******** ******* Center, ** ******* ******** ******.
Camera ********: ********* *** ******** ***** ****
*** ********* **-*********-*/** ****** **** ** Ambarella ***-*-**-** ****** ** ****, * common *** *** ***** **** ******.
*******, ** **** ******** ********* ************ ********** **** (***), ****** **** other **/******** ****** ***** **** ******** only * ****** ****. **** ** Movidius' ******* **********, ******** ** **** (see *** ******:***** ******** ******* ***** ************ ******).
Versions ****
*** ********* ******** **** **** ****** testing:
- ********* **-*********-*/**: **.*.** ***** ******
- **-*******-**/**/**: **.*.** ***** ******
*** ****, ** * *** ** a ****** ** **** ******* ** production, * ***** ****** *** **** checked *** "*******" ********* ***. ****'* exactly *** *******. ** ** *** checked **** ****** ********* ** **** scene, ** ***** **** ****** * significant ****** ** ********. **** ****** aren't ******** ** ***.
** **** ***** ** ** ******* that **** *** *** ***** *** kind ** **** ****** **************, **** commonly **** ***** **** ***/*** *********. Kind ** ******* ***** ***** *** of *** ******** ****, **** **** test **** ** *** **** ** be ***** ********* *** ************.
**'** ***** * ******* ********* ******** to *********'* ******** ********* *** ***:
Vs. ********* ******** ********* *** ***
******** ***********'* ******** ********* *********, ******** *********** ** ******* ******, rejecting **** ***** ****** **** **** sources, ********* ******* *** ******, **** alarms ** *******, ***. *********** ** drastically ****** **** ******** ***, ***** alerted ** *********** *** ***** ******, including *** *** ******* ** *** common ***** ***** ******* *****, ******* discriminating ******* ******, ********, ** ***** objects.
* *** ******* *** ******* *********** in ********* *** ******* ******* ******* compared ** *** ******** **********. **** review, ***** ***!
* **** **** **** ******* *** detection ********** ** *** ****** ** versions
************* ******, *** *** ******* * primarily *** **** *** **** **** a *.*.** ********. ** ***** *** current *** *** ** *.*.*, *** the **** ******** ******* ***** ** 5.6.0. ***** *** *** ***** ********* models * *******. (*.*.* ** **** available *** **** ***** ******* ******* Accusense.)
******* **** ********* ********** ** *****...*******, people ******* ******* *** ***** **** missed **********. ** ******** ******* ******** this *********.
******: **** *****?
******: *** ** *****, ****** *** to ** * ***. *********, *******...
*** *****. *** **** ** *** want ** *** ****, *** ** you **** **** ***** ****** ** you ***** *********** ** ** **, or ***? **** ******* ** **** might **** *** ****** ***** **** moving *******.
**** * ****** **** * *** adjust *********** *** ********* ** *** and *** ** ***** ***** * was *******, **** ****** ** ** max ******'* ******* ** *****, ** did *** **** **** * ******** box ****** **.
...*** ** ***** ***** * *** running, **** ****** ** ** *** wouldn't ******* ** *****, ** *** not **** **** * ******** *** around **.
***: *** *** *** ****?
**: *** ****?
SoC: If I knew “what”, I wouldn’t need you. Something just went flying by...
**: *** **** **** ******* ****** of *****. * **** ** ****. It’s **** ***. ******, *** ** frame, *** ** ****, *****?
***: **** *’* ***** **** ** alert ******, *** **** ****** ** notified!
**: *** **** *** *** ***** to *** ** ** **** ****, a *******, * ***, ***** * sand *****?
***: ***, ***’** ***** ***** *** getting *****. *** *** *** *****?
AI: Yeah, keep your trap shut. Remember, if faced with the choice of false negative vs. false positive, ALWAYS choose false negative. If you missed it, they probably will too; nobody’s the wiser. On the other hand, calling out “heavy snow” in July will just put us on the cover of IPVM, again!
***: ***’** ** *****!
****** *** *** *********** ***.
** **** - *** ** *** common ***** ******** ** ******* ** night. ** *** ***** ****** *** how ** **** *** **** ** would ** **** *******.
******, **** *** ******* ********* *****.
***** ** *** ***** * ** not **********.
********* *** ********, *****?
*** ***** ******** **** ** ** tuned *** ********* **********.
*** *** **** *** ** * good *** ** ******* **?
** ** *** ****** *** *** of *** *** ****?
*** *** **** *** ** * good *** ** ******* **?
** ***** ********** **** ********* ********** throughout **** **** *** ************ ***** feedback *** ***********. ***** **** ****** the ****** **** *******.
******. * ****** *** ** *** same **** ***** *************, ** *** do ****** ******** **** ***** ** scene.
*** *******, **'** ***** *** ******** to *** ******. ** ** ***** tests, ** ******* *** ** ********* for ******* *** *********** ** ********** the *********. ** **** ******* *** own ******** ******* ** *** **** effects **** **** (** **** ****, minimal ** ****).
Hikvision ********
**** **** ********* *** ***** *** opportunity ** ****** ******* ****** ******* but *** *** ***** *** *********** for ******** ******* ** ******* *********. They ******* *** **** ******** ******* worked **** ** *********** **** ******** NVRs, *** ** *** ***** ** saw ** ********** ** *********** ******* Acusense *******' ** ***** ********* *** Acusense ******* **** ** *********** **** the ***.
Settings ******* ****** ** *** ******
** *** ******, ** ******** ******** to *** ** ********* *********** ********, but ***** **** ******* ******** ******** the **** *******. ** *********'* ********* analytics, ********* ********, *** **** ******** adjustments *** ********* (**** ****** ** in ***** ****** **********) *** *********** ("size ** *** ****** ***** *** trigger *****"). ******* ** ***** ******** eliminated ***** ******.
** *****,
** ************* ** ***'* ******** ** you *** *** ***** ********* **-***** the ****** ** **** ** ***** into *** ***. *** *** **** work *** ********* ** ******** ****** feeds, ***** ******** ******* **** ******* the ********* ** ***-******** **** ** just ******* ******* ** ***** ************.
** **** *******? ** ** *'* thinking ** ******* ** *** ** the ******** **** ** ******* *** one ** **** ** *** **** which *** ******* ** ******* **** the ***** *********. ********?
***, ***, **** ** *******, *** Acusense *** ********* **** **** **** third ***** ****** ******* *** ******* people *** ******* ******. *** ******** cameras **** **** ***** ************* ** non-Acusense **** *** ***'* ******* *** classification.
*****,
* ******* ***** ***** ** * typo ** *** "Vs. **** **" section of the article. You mentioned "Axis Perimeter Defender" should that be Acusense instead?
******,
You **** :
Detection ********
** *** *****, ******** ******** ******** at ~***' ********, ~** *** ** the *** **-*********-*/** ****** ** ******
********:
*** ****** ******* *** **** **** in *** ****? *.***, *** ** 8mm?
*** ******* ***** *** ***** *.*.* or *.*.* ********? **** **** ****** be * **** ** **** *** Hik ** ** ****** ***** ***** Analytics+.
*** ***** *+ ****** * *** also ******* *** ****** ****** (*/*.*), where *** ** ***** *** ******* size ******* ** ***** *** ** lower *** ******.
** ** ***** *** **** **** standard ****** ****** (***) ** ******** when *** ******** ********* *** ******* resulting ** ********** *********? * ****** the **-********* **** ** ***** **** out *** ***** *** **** **** VMD *** ******** ******* **************.
***, * ******* ******** ** *** DS-2CD2346G1-I/SL ** **.*.* ***** ****** *** I ***'* *** ** ****** *** VMD, **** **** * **** *** smart ****** ** *** ******** ********.
****'* *** ***. **'* *** * huge **** ** *** **** **** of *** ***** ***** ** ***, but *** ********** ********** ***** * MicroSD ***** ** *******.
** ***** **** ***** ** * way ** **** *** *** ********* for ******* ** ***** ** ** be **** * ***** **** ***** be ******* *** ***** ** * small ******* *****. *******, **** ***** leads ** ** ******* **** *** be ***** *** **** ** ** object ** *** ***** ** ************* between ***** ** ****?