US Border RVSS / Video Analytics System Examined

Published Apr 21, 2017 12:11 PM

US Customs and Border Protection has been rolling out a video analytics-based detection system along the US/Mexico border, with detection ranges reaching up to 7 miles. Deployments began in 2015, and were recently declared to be at Full Operating Capability after 2 years of testing

IPVM spoke with an executive of the company behind the analytics, PureTech, about the project, operational requirements of the analytics, and what led to their selection for this project. Details of the conversation are in this report.

RVSS ******* **********

*** ****** ***** ************ ****** (****) used ***** ******** ** *** ** border **** **** ****** *** ****** employs ****** ** ********* ******* ** to * *****, ******** ** ******* observation *** ********* ** ************ ****** crossings **** *** **. * ********** update *** ********* ** ~**** ** select *** ******** ** ****** *** agents **** ********* *** ******** ************ border *********. ******* ******** **** ********* by *** *** ***** ********** ******* Dynamics, ***** ********'* ********* **** ******* to ** **** ********** ** *** project. *********** ***** ** ****, ***** products **** ******** ** **** ***** in ******* *** ***** *** ** extended **** *****, ***** ****** ~* years. ** ***** **** *** ***** systems **** ****** ** ** ** "Full ********* **********", ******* **** **** essentially *********** ** ******** ** *** live **********. **** *** ********** ***** 1 ** *** *******, *** ******* roughly ** ***** *********.

***** * ** *** *******, ** ***** ** ********** ~** tower ********* ** *** ******** *****/*** *****. **** ******* ** *** ******** to ******** *** ********, *** **** to ******* **** *** ************/******* ** the ******** ******* ** *** ********** locations.

Overview ** * ******* *****

*** ********** ****** *** ******* ** a ********* *****, *** * ***** tower **** **** *********. *** ****** house ******* *** ************** *********, ***** with ******* ** *** *** ********* software, ** *********** *** ***** ** a ******* ******** ** *** ********* due ** ********* ***********, *** ******** does *** **** ** ******** ******* of ***** *********. ****** *** ******* fixed/permanently *******, ****** *** ******* **** have * ****** ** *********** ****** towers **** *** ** ***** ** different ****** ********* ** *********. 

**** ***** ********* ******** ******** ******* cameras- *********, **** ****** ******** ** General ********, ** **** ** **** to ******* ********** ****** ******* ** targets ********. * *******-***** ****** ************'* ********* ********, *** *********/****-***** ******** ******** ***** data **** ** ******* ******* *********.

Command *******

****** **** ********* ******* ** *** individual ****** *** **** **** ** centralized ******* *********, ***** ***********'* *** (****** ***** *****), ***** ***** ********* images ** *** ******* ****** ********* a ******, ********** ****** ***** **** the ***, *** * *** ******* where *** ****** ** ******* ** real ****. ******** ******** ****** **** range, *** ***** ** ** ******** within ******* ******, ******** ****** ** the ***** ** ******** ** *** location ** ********* *** ************ ******* crossing *** ******.

Project ************

*** ******** ******* **** ***** ****** (but *** ********) ******* ** ****** up ** * *****, *** ******* some ****** ** ******* ******* **********. GPS *********** ** ******** ******* **** also ******** ** ***** ** **** live *** ******* *** **** ***** agents *********** ** *** *** ************ along *** ******. ********* ** ********, their ****** *** *** **** *** evaluated **** *** **** ** **** the ******** ******** ************ ** *** CBP.

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

********* ** ********, ***** ********* *** designed ** ** ****** ****-***, *** covering ************* **** ****** (** **** cases ** ** * *****) ******** additional ****** ** ******* *** ******** accuracy *** ******** *** *** ******. The ******* **** *** ********* ***** need ** ** ***** ** *** right ********* *** ******** ** ******* a ***** ** **** **** ** narrow ****** *** ********* ********* *** **** accounts *** ***** **** ** ********** from **** ** ****. ******* ******** was *********** *** **** **** ** the *******, ***** ***** **** ******* hours ** * ***, ********* ** the ****. **** *** ******* *** aimed/focused, *** ********* ******* ******** **** a ****** ****** ** *** ***** of **** ** **** *********** *** be **** ** ******* *** ********* changes *** ***** ******* **** *** specific ** **** ******. **** ******* takes * *********** ~*/* *** ** complete ** *******, ********* ** ********, though *** ****** ********* ** ********* by ******* ********.

7 **** ********* *****

** ***** ** ******* * * mile ********* ***** ******* ******** *** PureTech **** ******* **** ******* **** an ~* ****** ********** ******* *****. These ******* ** *** ******* * miles ** ********** ********, ****** **** detect ******* **** ******* * ***** of ******* **** * ******** ** 5-7 ***** **** *** ******/*****. ********** cameras *** **** ** ***** *** area **** **** *** ****** ** in *** ***** ******* *** ***** and *** ***** ******** *** **** camera ******.

Analytics **** ********* ******

*** **** ****** **** ****** ** the ***** ********* ** ****** *** identify *******, ** ************ ************ (****** radar, ****** *****, ******* *********, ***.) are ******** ** ***** *********.

Comments (4)
MC
Marty Calhoun
Apr 21, 2017
IPVMU Certified

Interesting article, thank you for the information

(2)
UE
Undisclosed End User #1
Apr 21, 2017

I actually agree with Marty!!!

(1)
(1)
(4)
Avatar
Michael Gonzalez
Apr 21, 2017
Confidential

Pretty impressive detection range, though I'm not sure why target classification wasn't deemed important enough to include in the bid specs. The end state of this project means there will be hundreds of miles being covered by this system, target classification would rule out a lot of nuisance alarms. I guess they'll figure that out when they get there...

(1)
Avatar
Bob Kusche
Apr 22, 2017

I think that while the PureTech solution is impressive with its distance capabilities, the credit for that needs to go to Flir more than PureTech.  But I agree that lack of target classification is a major weakness.

I attended the Border and Security show in San Antonio last week, and a couple of things stood out to me: PureTech wasn't there (I suppose they were too busy digesting their current contract?), and the need for range vs. detail vs. short-range and more portable solutions is very apparent.

A 1% field of view gives us the same weakness we have always had with PTZ cameras. If we are serious about locking up the border, a hybrid approach of solutions will be the final answer.

Personally, I feel that the minimum capabilities of the solution must include:

  -- Very long range cameras (like the cameras mentioned in this article)

  -- Analytics (based on sensors from a combination of long-range video and shorter-range ground-based-radar with better classification capability)

  -- Drones (for faster dispatch, especially over rough terrain) that can store and, push video to a central VMS where infrastructure allows, and can self-charge

  -- Portable systems (for faster setup -- less than two hours -- by a single person with limited/no need for re-calibration over time)

  -- An industry-standard, open-architected, VMS

  -- Specialty cameras (4k, very low light, etc) where indicated

  -- Oh, and a humungous wall where feasible!

 

(3)
(1)