Axis Object Analytics Time In Area Tested
Axis has released Time In Area analytics as part of their AI-based Object Analytics, intended to detect loitering of people and vehicles, but how well does it work?
We tested the new Time in Area feature outdoors and indoors for seven days, examining:
- How reliably does it alert on humans loitering in the region of interest?
- How reliably does it alert on vehicles in the scene?
- Does it accurately classify cars, trucks, vans, and other vehicles?
- Does it accurately alert on subjects in groups?
- How do subjects approaching the camera vs. crossing the scene affect performance?
- How complex is time in area to configure?
Note: This report focuses on Axis' new Time In Area feature. For more on Axis Object Analytics' general performance, see our full test results: Axis AI Object Analytics Tested.
Configuration ********
** ****** * **** *** *** time ** **** *********, ***** ****** on *** **** ** **** ******* and ***** *** ******* ****. **** is ******* ** ******* *** ****** from * ** *,*** (* *****). No ***** ************* ** ********, ***** is ***** ** *** ***** *****:
************
**** ** **** ** ********* ** all ******* ********** **** ****** ********* at ** ********** **** ******** ** firmware **.*, ********* ** **** ****. Previous ******** ******** **** ******* *** showed ** ** **** ** *** web **.
Executive *******
***** ***'* **** ** **** ********* accurately ******* ** ****** *** ******** loitering ** ******* ** ********, ********* when ********* **********, *** *** *** false ***** ** ****, ******, ** other ****** *******, *** ********* ************ dropped ******** ** *** ** **** people ** ******** **** ********* ** a ***** *** ******** *****, ********* the **** ** **** *****.
People/Vehicles, *** *******
**** ** **** *** ** **** to ****** ****** *** ********, *** not ***** *******. *** *******, ***** it *** ****** * *** ****** over **** ** * ******* **** or ****** ********* ** ** ************ area, ** **** *** ****** * backpack ** ****** ******* ** ** area. **** *** ************ ******** ** as ** "****** **** ******" ** other ********* *** *** ******* ********.
**** ****** ********* ********* *** ** set ** ****** "*** ******," *** this ******** ****** *** ******* *********, does *** ******* **** ** **** settings, *** **** *** ************ ****** and ******** ****.
No *****/******** *********
************, ***** ** ** ****** ** alert **** **** * ******** ****** of ******* *** ********, *.*., **** alert **** *** ** **** ****** remain ** * ****. **** ****** Analytics *** ********** ** "********* ** area" ********, *** ** ** ********* in **** *** **** *** ******* time ** ****.
Competitive **********
****' **** ** **** ** *** only **-***** ********* ******** ** **** tested. *******, ******* *********** ******* *****, including******** *** (***** *****),***** ***,****** ******* **, *** ******.
** *** ****, ** ****** ******* non-AI ********* *********, **** ** **** Guard ***** *** ****** *** *********. We ***** **** **** ******** **** AOA's **** ** **** ***** ***-** analytics **** ************ ******* ******** ******* the *****.
Accurate ***** *** ******* *********
** *** *******, *** ********** ******* on ***** ******** ********* ** *** field ** ****, ********* **** ******** past **** *****, ********* ******** ** otherwise **********.
*** *** ***** *** ****** *****, changing **** ****** ** ***, ** they ****** ** *** **** ***** time ******* *** *** ***** ** generated.
********* *** ******** ** *** ***** (1-5 ***) *** ** *** **** with ** **:
************, ******** **** ********** ******** ********, accurately *********** ****, ******, ****, *** other ********. *** *******, *** ***** below *** ********* ** * *** parked ** ** **** ***** **** trucks *** **** ****** ** *** more **** *** *******:
Humans ******** ***** ********
*******, *** ********** ******** * ***** subject ***** ** ******** ********* ********. Shown *****, *** ******** ********* ********* of *** ******* ** ** ******* behind ***** ******** *** ***** ****/***-****** of *** ****:
Subjects **** ** ******/******* *** *********
**** ******** ***** ** * ********* group *** ******* *****, *** ******** frequently ******* ********* ** *** ** more ******** ** **** ******, ********** the **** ** **** ***** **** detected *****:
***** **** ** * ******* ********* issue, *** ***** ******** ** *** group ***** ********* *** **** ** area ***** ** *** ****** ** interest, *** ***** ***** **** ******* were **-******** *** ********* ****** **** the ***** *******.
Humans ***********/**********
** ******* ****** ********* ****, *** ******** ********** ******* ******** of ***** ******** ******** *********** ** moving **** **** *** ******, ********-**, instead ** ******** *** *** ** moving ** ** *****. *******, ** found **** **** ** **** ***** reliably ****** ***** ********, *********** *** countdown ***** **** ** *** ******* was **** *** * *** ******.
Strong ********** ** ****
** ** *** ***** ***** ** AOA, ** ***** ** ***** **** In **** ****** ****** ** ***** rain ** ******** ** *** ******'* dome. ** ***** *****, ******** *** glare ****** ** ****** ********* **** them *** *** ******** ** *******:
Person ********* ******** ** ~** ***
**** *** **** *****-*** ** *** widest ***** ** **** (***°), ***** detection *** ******** ** ~***'/~** ***:
**** ****** ********* ******** * ***** to ****** **** ** *** **, showing *** *********** ******* ****** **** for ******** *********, ******* ******** **** we ***** ** *******.
VMS ***********
**** ** **** **** *** **** alerts ** ***** *** ********* ****** and ** ********** **** *** **** VMSes ** *** *******:
*** **** ******* ** ******** *********** issues, *** *********** *** ****.
Versions ****
*** ********* ******** *** ******** ******** were **** ** *******:
- *** *******: **.*.**
****** *** ******** **** ** *** attention, **** *** ** ***** **** making *** ******. ** **** **** to ********* **** ****** **** ******** via ***** ******. * ******* *** chart ** ******* ****.
**** *** ** * **** ********, but **** ***** ******* *** *** events ************ ******. ***** ******* *?
***, **** *** ******* * **********. We **** **** ***** **** ** our******** ****.
*** ***** **** ** *** ********* areas/zones ***** **** ******** ** *** existing ********* ******** (** ***** *** my *** *****).
*** *** *** ******** *****, *** not ** *** **** **** (***** they **** * ********). *** ***** have ** **** ********* ********* *** each ****. * *** **** ** our ******* - * **** * scenarios ***** **** ******* ******** ***** in *** **** ** *** ******** and **** *** ******* ****-****** ****, so * ***** **** *** **** was **********. * ******* *** *** make ** *********. * *** ** 10 **** *** ************* ********:
*** *** *** ******* *** ****** to ***. **** ******* ** ***** drivers? *** *** **** *** ***** events ** *** ** *** *** receive *** ******?