Uniview's Chile Bus Temperature Screening Is Faulty

By Robert Wren Gordon, Published Nov 13, 2020, 11:28am EST

[Por favor haga clic aquí para leer este artículo en español.]

While Uniview has touted its bus-mounted fever screening solution as a "success case", IPVM's testing showed that false readings are a fundamental problem, with Uniview admitting that they will be affected by this.

IPVM Image

Now, 900 Chilean miners and other bus riders that depend on these systems are at risk.

This is the seventh in a series of posts about faulty fever detection deployments in Latin America. For previous coverage, please see:

  1. 🇧🇷 Hikvision Fever Cameras Endanger Public In São Paulo, Brazil Malls
  2. 🇲🇽 Faulty Hikvision Fever Cam Setup at Mexico City Basilica and Cathedral
  3. 🇨🇱 Dangerous Hikvision Fever Camera Showcased by Chilean City
  4. 🇨🇴 Faulty Hikvision Cali Colombia Fever Camera Implementation
  5. 🇦🇷 Dahua Buenos Aires Bus Screening Violates IEC Standards and Dahua's Own Instructions
  6. 🇨🇴 Colombia's President Promotes Bad Hikvision Fever Camera Setup

Chilean Mine Case

Uniview released two press releases, one in English and a second in Spanish, in late September touting the deployment of Uniview's OTC-513 Intelligent Standing Pole-mounted Measuring Instrument at Chile's Minera Valle Central (MVC) as a "success case" and safety enhancement:

IPVM Image

The MVC is located around 100 km (62 mi.) south of Santiago, Chile's capital in Cachapoal Province. The mine is owned by Vancouver-based Amerigo Resources Ltd. and has been contracted by Codelco, a Chilean SOE, to recover copper and molybdenum from tailings.

Uniview's press release describes that the MVC has 900 workers and was looking for a way to accurately measure temperatures on its eight buses:

Join IPVM Newsletter?

IPVM is the #1 authority in video surveillance news, in-depth tests, and training courses. Get emails, once a day, Monday to Friday.

Join IPVM Newsletter?

IPVM is the #1 authority in video surveillance news, in-depth tests, and training courses. Get emails, once a day, Monday to Friday.

IPVM Image

As a solution to the MVC's problem, Uniview touts the OTC-513, highlighting its "accuracy of 0.1°C" (0.18°F):

IPVM Image

The press release later clearly states that "the OTC-513 is installed to the bus entrance," providing this photo:

IPVM Image

The both the English and Spanish versions of the press release then end with quotes from MVC employees praising the deployment, with one MVC employee stating that the mine was so happy with the device, that they expanded the Uniview deployment, confirming that “two Temperature Control devices were implemented in two dining halls.”

IPVM Image

IPVM Testing

IPVM bought and tested the OTC-513, the same model used by the MVC. Beyond the fundamental issues we found, a particular problem was temperature screening with people outdoors. In particular, variances in outdoor temperatures caused significant variations in Uniview reported temperatures.

For example, in the test shown below, a subject was outside for five minutes in a light, long-sleeved shirt and the reported wrist temperature dropped 4.28°C (7.7°F):

IPVM Image

By comparison, the difference between 'normal' and 'fever' is less than 1.1°C (2°F), thus significant swings in temperature are a major concern.

Uniview Response / Admits Problems

Uniview responded to our test results, arguing that this is 'unavoidable' and that 'fast screening' is undermined by changes in temperature, such as being outdoors:

At present, all devices in the market that can support fast screening without contact will indeed be affected by the ambient temperature. This is unavoidable because what the device measurement [sic] is body surface temperature.

The manufacturer encouraged people to wait five minutes and test again:

There are warning [sic] on the device to avoid low body temperature: if the device detects low temperature (the default threshold is lower than 35.5°C (95.9°F)), it will prompt Low temperature. Then, the subject should wait indoors for about 5 minutes for test again. Usually the body surface temperature will return to normal at this time, and the measurement result will be accurate.

This is particularly not realistic for boarding a bus, since either the bus has to wait with presumably the person inside or the person needs to wait for the next bus and hope they can somehow warm the wrist up to a level that will pass the Uniview system.

Uniview concluded by contending that if a person is wearing thicker clothes to cover the wrist that the temperature reading should be less affected:

[W]hat we talked “the wrist is less affected by environmental factors”, is compared with forehead temperature. Because in low temperature environment, people will wear the thicker clothes, wrist usually will be protected more than forehead.

Presuming that users will wear clothing that sufficiently covers the wrist without allowing it to get too cold or too warm exposes the fragility of this type of screening.

To Be Continued...

This is part of an ongoing series highlighting bad fever detection deployments throughout Latin America. Fever detection cameras and tablets are expensive solutions that, when improperly deployed, not only do not address the underlying health emergency but could make it worse by providing a false sense of safety. IPVM will continue to expand its coverage of fever detection deployments and advocate for their proper use. If you know of any flawed implementations that we should profile, please post in the comments or email us at info@ipvm.com.

El monitoreo de temperatura de Uniview en buses en Chile es defectuoso

Mientras que Uniview ha promocionado su solución de monitoreo de temperatura en buses como un “caso de éxito”, el testeo de IPVM detectó que las lecturas falsas son un problema fundamental, mientras Uniview admitió que se verán afectados con esto.

IPVM Image

Ahora, 900 mineros chilenos y otros pasajeros de los buses que dependen de este sistema de monitoreo se encuentran en riesgo.

Este reportaje es el séptimo en una serie de artículos sobre mecanismos de detección de fiebre defectuosos en América Latina. Para acceder a estos, por favor visite:

  1. 🇧🇷 Camaras de Hikvision ponen en peligro al público de los malls en São Paulo, Brasil
  2. 🇲🇽 Instalación defectuosa de cámaras de fiebre de Hikvision en una Basílica y una Catedral de Ciudad de México
  3. 🇨🇱 Peligrosa cámara de fiebre de Hikvision instalada en ciudad chilena
  4. 🇨🇴 Cámara de fiebre defectuosa de Hikvision implementada en Cali, Colombia
  5. 🇦🇷 Detector de temperatura de Dahua en buses de Buenos Aires viola sus propias instrucciones y los estándares del IEC
  6. 🇨🇴 Presidente de Colombia promueve instalación de cámaras de fiebre defectuosas de Hikvision

Caso de la minera chilena

A finales de septiembre, Uniview entregó dos comunicados de prensa, uno en inglés y otro en español, promocionando el despliegue del OTC-513 equipo de medición inteligente montado en poste de piso en la Minera Valle Central (MVC) en Chile, como un “caso exitoso” y que supone un mejoramiento de seguridad:

El producto rastreador de temperatura de UNV asegura a los pasajeros de buses en Chile

IPVM Image

La mina MVC se encuentra a unos 100 kilómetros (62 millas) al sur de la capital de Santiago, en la provincia de Cachapoal. La mina es propiedad de Amerigo Resources Ltd., una firma con sede en Vancouver que fue contratada por Codelco, empresa estatal chilena, para recuperar cobre y molibdeno desde los relaves mineros.

El comunicado de prensa de Uniview describe que MVC tiene 900 trabajadores y que está buscando una manera de medir la temperatura en sus ocho buses:

IPVM Image

Como solución al problema de MVC, Uniview promocionó el OTC-513, destacando su exactitud.

IPVM Image

Luego, el comunicado de prensa claramente dice que “el OTC-513 es instalado en la entrada del bus”, entregando esta foto:

IPVM Image

Tanto la versión en inglés como la versión en español del comunicado de prensa terminan con citas de los empleados de MVC, alabando el despliegue, con unos de los empleados de MVC declarando que la mina estaba muy feliz con el dispositivo y que habían expandido el despliegue de Uniview, confirmando que dos dispositivos de control de temperatura fueron implementados en “dos comedores”:

IPVM Image

Testeo de IPVM

IPVM compró y testeó el OTC-53, el mismo modelo usado por MVC. Más allá de los problemas fundamentales que encontramos, un problema particular detectado fue en el monitoreo de temperatura de personas que se encuentran al aire libre. En particular, las variaciones en la temperatura en exteriores causa una variación significativa en las temperaturas reportadas por Uniview.

Por ejemplo, en la prueba mostrada abajo, un sujeto con polera manga larga delgada estuvo en el exterior durante cinco minutos y la temperatura de muñeca reportada bajó 4,28º Celsius (7,7ºF):

IPVM Image

En comparación, la diferencia entre “normal” y “fiebre” es menos de 1,1ºC (2ºF), es decir, los cambios significativos de temperatura son una gran preocupación.

La respuesta de Uniview / Admiten problemas

Uniview respondió a nuestros resultados de testeo, argumentando que el problema es inevitable y que los monitoreos rápidos están determinados por cambios en la temperatura, como los provocados por estar expuestos al exterior:

En el presente, todos los dispositivos que se encuentran en el mercado que realizan monitoreos rápidos sin contacto, efectivamente se ven afectados por la temperatura ambiental. Esto es inevitable ya que los dispositivos miden la temperatura superficial.

El fabricante del dispositivo fomentó a las personas a esperar cinco minutos y volver a medir la temperatura:

En el dispositivo hay advertencias para evitar las bajas temperaturas corporales. Si el artefacto detecta bajas temperaturas (el umbral determinado es menor a 35,5ºC), Entonces, el sujeto deberá esperar al interior por aproximadamente cinco minutos para volver a medir la temperatura. Usualmente la temperatura superficial corporal volverá a su normalidad y el resultado de la medida será precisa.

Esto no es particularmente realista al momento de subir a un bus, puesto que este debería esperar a que la persona se encuentre dentro o el sujeto tiene que esperar durante cinco minutos para el siguiente bus y ojalá pueda calentar su muñeca de alguna forma, hasta llegar a un nivel que el sistema de Uniview aprobará.

En su respuesta, Uniview concluyó que si una persona está usando ropa más gruesa que cubra su muñeca, la lectura de temperatura debería verse menos afectada:

Dijimos que “las muñecas se ven menos afectadas por factores ambientales”, en comparación con la temperatura de la frente. Esto debido a que ante menores temperaturas ambientales, las personas usarán ropa más gruesa y las muñecas usualmente se encontrarán más protegidas que la frente.

Presumir que los usuarios usarán ropa que efectivamente cubrirán las muñecas sin permitir que estas se vuelvan o muy heladas o muy calientes, exponen la fragilidad de este tipo de monitoreo.

Continuará

Este reportaje es parte de una investigación en marcha sobre malos mecanismos de detección de fiebre a lo largo de Latinoamérica. Las cámaras y tablets de detección de fiebre son soluciones caras que cuando son implementadas erróneamente, no solo enfrentan de manera incorrecta la emergencia sanitaria, sino que además podrían empeorarla al entregar una falsa sensación de seguridad. IPVM continuará su cobertura sobre despliegues de detección de fiebre y abogará por el buen uso de estos dispositivos. Si sabes de alguna implementación defectuosa que podríamos investigar, por favor déjalo en los comentarios o envíanos un correo a info@ipvm.com.

Artículo traducido al español para IPVM por Camila Higuera.

2 reports cite this report:

Hikvision Co-Sponsored Smart City Expo Santiago de Chile Examined on Dec 15, 2020
IPVM attended both days of the Hikvision-co-sponsored Smart City Expo...
Non-PRC Manufacturers Block Dahua From Chile Event on Nov 30, 2020
When Dahua registered for an industry event in Chile, several non-PRC...

Comments (4)

Only IPVM Members may comment. Login or Join.

This is IPVM's first bilingual English/Spanish post! 🥳 Thanks to Chilean journalist Camila Higuera for the translation!

As some of our Latin America-based readers already know, this is just the start of IPVM's ongoing efforts to expand our reach in Latin America and make IPVM content more accessible to industry professionals, journalists, and members of the public with limited knowledge of English. We look forward to publishing more content in Spanish and Portuguese and to expanding our reach in Mexico, the Caribbean, and Central and South America!

If you have any suggestions for how we can improve or any story ideas, feel free to reach out to me (in English, Spanish, or Portuguese) at rgordon@ipvm.com.

Thank you to everyone who's made time to meet and/or speak with me so far, I share your feedback with John and Don and am looking forward to providing more in-depth LatAm coverage in 2021 and beyond!

Agree: 4
Disagree
Informative: 2
Unhelpful
Funny

Actually yes, this use case is not very clever and I'd blame the SI and the MVC for this more, then UNV although it's UNV's mistake to put it into the official marketing. I'd not do it.

Agree: 1
Disagree
Informative
Unhelpful
Funny

I'd blame the SI and the MVC for this more, then UNV although it's UNV's mistake to put it into the official marketing

Petr, are you going to blame the end user more than the manufacturer who literally turned this into a marketing campaign? How is the end user (MVC) supposed to know that this is a faulty approach if the manufacturer of it is literally touting it globally?

Agree
Disagree
Informative: 1
Unhelpful
Funny

John, maybe I'm wrong, but in this case, the first was a project and the success story itself afterwards. End User is responsible for the tender conditions, this is where I see the mistake. It was easy to avoid this situation for the MVC.

This product is originally designed for indoor windless applications as is written in the specs. We can implicate it's designed for the office purposes not for the transportation.

We are UNV distributor in the Czech Republic, we know UNV marketing pretty well. I'd say there is no marketing, it's always a photo of a product and some useless text. From time to time comes sucess stories. I'd never say UNV promoted this as a solution for transportation => SI and MVC's thing, not UNV's.

Agree
Disagree
Informative
Unhelpful
Funny
Loading Related Reports