Digital Watchdog MEGApix Ai Analytics Tested
In 2023, IPVM testing found DW analytics inferior, but now Digital Watchdog has released its MEGApix Ai camera line claiming "AI Deep Learning object detection" but how well does it work?
We bought and tested the DWC-XSDE08Mi, which uses their MEGApix Ai analytics, examining:
- *** ** ******* ** ******* ******* to ***** **** *********?
- *** **** ** ******* ** *** performers **** ****, ********, ** *****?
- **** ******* ******* ***** ********* ** common ******* ** ***** ****** **** as ******* *******, *******, ** ****** vehicles?
- *** ** ****** ***** ********* **** the ******* ******* **** ** ******* or ****?
- *** ******** *** ****** *** ******* classifications?
- *** ****** ********** ******* ****** *** scene?
- **** *** *** ****** ******** **?
- **** ***** **** ** ********* ****?
*******
*** ***-********, **** ******* ** *********, is ********* ** ** **** ** ~$600, **** ** ********** ********* **** for *** *********. ** **********, *** Axis *****-***, ***** **** ******** ******* features ********* ** **********, ********* **********, suitability *** ******* ***, *** ****** Analytics, ******* *** ~$****, ******* ~$*** more **** *** ***-********.
*******
******** ******* ** [******* ********'* ************** of "**", *** ****'*] *** ******** in ********** *********** ******* ******** ** their ******** ******* *** *********, ** still ***** ** ** ******* ** below-average ********* **** ******** ******* *** performing ********* **** ***** **** ****, Avigilon, ** *****.
** *** *******, ******* ******** ******* Ai *** ** ***** ********* ********* by *******, ****, ****** *******, *** changes ** ********. *******, ** ************ generated ***** ********* ** *******, ****** vehicles *********** ** **********, ****** ********** objects ** *** *****, *** ****** repeated ****** *** * ****** ********** present ** *** ***** *** ** poor ******** *** ********.
Improvement **** **** ********* ****** ***** ******* ******** ** *** **********
**** ******* ***** ******* ** ** be ** *********** **** ***** **** generation ******* *** *********. ** ** longer ******** **** ***** ********* ** sources **** ** ******* ******* ** shadows ** *** *****, *** ** no ****** ***** ***** "*******" *************** on ******** ** ********* *******. *** setup ******* *** **** **** **********, and *********** ** ** ****** ******** before ***** **. *******, **** ******* IVA, ********** **** ** ******** ** reduce ***** ********* ** ******* *** inanimate ******* ** *** *****.
**** ******** ** *** ********** **** as ****, ********, ** *****, ** is ****** ******* *** ** *** many ****** **** ***** ********* ** animals, ********* *******, *** ************ ** parked ******** ** *****. ******* ** also ******** ********* ******** ** ****** false *********, ***** *** ********** *** not **** ****** ***** *** ******* setup.
False ********* ** ******* ** *** *****
******* ******** *** *** ****** *********** animals ** ****** ** *** *****, whether ** *** *** ****, ** the ********** ** *** *****, ***, or ***** ***** *** ****** **, animals **** ********** ** *****, *** an ***** *** ********.
Fluctuating ********** ******* ** ******** ****** ** ****** ****** ** ~** ***
****** *******, *** ******* ** ******'* confidence ** *********** *********** ****** * scene ********* ********** ******* ************* **% to **% **** *** ***** ******* is ** ***** ** ***. **** confidence ***** ***** ***** *** ******* threshold, ********* ** **** *** ***** on * ******. *******, **** *** pixel ******* ********* ** ****** ** PPF, *** ******'* ********** *****, ******* between ************* **% ** **%. **** reduced ********** ******** ******** ****** *** a ****** ********** ** **** ***** any *** ********* (**** *** ******* threshold ***** **%).
********* *** ******* ********* ********* ** 30% *** ********** *** ******* ******** threshold *** ***** *** ****** ** multiple-person ******, *** **** ****** ** fully **********. **** **** ********** *** initial ******** ********* **** ****** *** detection ******** ** *** ******.
Headlights ******* ***** ** ****** *******
** **** ****** (**), ** *** headlights ** * ******* ******* ********** a ****** *******, *** ****** ********* this ** ** ***** *** *** parked *******.
*******, ** * ******** ***** **** consistent ********, ******* ** ************ ******* moving ******** *** ******* ****** ********.
False ****** ** ********* *******
** *** *******, ***** ****** **** seen ** ********* ******* ** *** scene, **** ** * **** ** an ********** *******.
*******, ***** ****** *** ** ******* by ********** *** ******* ********* *** tracking.
Lower ********** ** ****** ********
******* ** ******** ******** *********** ******** with ********** *********** ******* ************* **% - **%. ** ********, **** * person ****** *******, *** *** ***** density ** ***** ** ***, *** system's ********** ******* **** **********, ******* from **% ** **%. ******* *** variations, *** ********** ****** *** ********* crawling *********** **** ***** **** ***** the ******* *********, ******** **** **** not ******.
People ******** ***** ********** ** *******
****** ********** ** ******* **** ********, though ********** ******* ** ~ **% - **% **** **** **** *** seen.
No ***** ****** ** ******* ******* ** ****** *******
****** *******, ** ******** ** ***** events ** ******* ******* ** *** shadows **** **** ** *** ******.
Poor ****** ** ******* ** ***** ****** ********* *** ****
** * ******** *****, ******* ** detected ****** ** ~** ***. *******, the ****** **** *** **** *** specified ** ***** ** ~***' *** PPF ****** ** ***** ********* ** ~40.
Setup ****** **** ******* ***, ** *********** ******
******* *** ********* ******** ***** *********** for ******* ***********, ************ *** ******'* height *** **** *****. ** ********, MEGApix ** ********** *** ***** ******* by **** ********* *** ******* ** a ******** ***. ******** ******* ** is ****** ** *********, **** ******* might ***** **** ********* *********** ** minimize ***** *********.
***** **** ***** ******* ***** ** easier **** ******* ***, ******** ** most **** *********, ** ** **** an ******* *******. *** ******** ***** for *** ********** **** **** ****** Analytics, ******** **, ***** *** *** no ****** ******* *** *********** *** only ******* * ******** *** *** choice ** *************** ** ***** ********* events.
VMS ***********
******* ** *** * ***** *** integration **** *** *** ******** ******* Center.
Versions ****
*** ********* ******** **** **** ****** testing.
- ******* ******** ***-********: ****.**.******.*.*
***, *** ****** ******** *****. *******, whenever * ********* ** *** *** camera ** ******** *** *****, ** would *** *** ** *** ***.
**** ** *******, * ***** ******** times ** *** ** *** ***** and ******* ** ***** **** *** options ******* ** ********* ***** *** need ** ****** *****, ***** ** does *** ****. * **** ****** the ****** **** *** *** ** make **** ** *** *********.
************, *** ****** ** *** ** the********* ****** ****(* ***'* **** ** *** ****** list ** **, *** **** ** search ** **** *************) *** ***.
**** **** ****** ******* *****? *** did *** *** ** ******* ** to ***?