23/01 2023
modern technology:
Surveillance cameras in Las Condes automatically generate crime prevention alerts
Safe Neighborhood Program
Almost 900 cameras in the community network are able to intelligently detect, support and reinforce the work of the Public Safety operational center.
Since 2020 and in line with the commitment to make crime prevention in Las Condes more and more effective, the Municipality has implemented an innovative technology of intelligent analytics for video surveillance of the municipality: it is a automated system where hundreds of “trained” cameras are capable of detecting potentially criminal situations.
In a network that currently includes more than 2200 cameras, the implementation of this analytical resource It already reaches 900 devices. And although outwardly similar to other less advanced surveillance devices, these new models are capable of automatically generating various types of alerts in real time, which are reviewed by operators of the Directorate of Security Operations Department, who evaluate in any case the eventual activation of the protocols.
For the mayor Daniela Peñaloza, one of the main objectives of this project is “the prevention of crime, such as robberies at home, vehicles on public roads or types of vehicles. Important in any innovation is to predict criminals and technology is essential in that there is also concrete evidence, high definition images, for the Prosecution”.
According to Jerav Rice, the official responsible for the project, called Protected neighborhoodthe operation of this intelligent discovery consists of several stages: “The first is related to the algorithms of deep learning, who are able to identify objects. The second is the alarm configuration, which is done through rules that seek to contextualize a target (wandering detection, crossing a wall, or slamming a gate, among others) with the image captured by a camera. And finally, that this configured alert conforms to the detection rules and execution conditions, and that it is received by security operators in a desk of direct operation”.
This set of techniques and algorithms machine learning Y deep learning are an important branch of studies artificial intelligenceand thanks to her Las Condes Analytics is able to identify people, types of motorized vehicles (such as cars, motorcycles, trucks and buses), bicycles, animals (dogs and cats), bags, cones, barriers, backpacks or waste containers.
Likewise, you can identify the colors of objects and determine the actions and poses of people. Efforts are also being made to detect persons with firearms, who throw objects and fights. In the same way, the flow of vehicles and individuals on the road can be analyzed, the license plates of cars to make a cross with those that are stolen or involved in crimes.
In terms of the reliability of the results provided by intelligent detection, the Department of Security Projects claims that “accurate events received by operators are between 90% and 95%. The idea is to leave a false positive rate for events so that there is a margin of interpretation for the detection patterns and this margin remains between 5% and 10%.”
Support for the analysis of post-crime events
The municipality’s surveillance cameras capture images every second through their high-resolution sensors and are connected to storage systems capable of storing video for weeks.
In the past, and with the requests of the police or the Prosecutor’s Office for the examination of audiovisual evidence, the huge volume of data could create obstacles, such as slowness in searching for specific events or details. An obvious example is the investigation of car crimes, especially when we try to know their patents, characteristics and escape routes. However, according to the coordinator of the Department of Public Safety’s Digital Evidence team, through new “analytics or computer vision,” images can be imported, processed, analyzed and understood in a much more efficient way.
“Large volumes of image data can be incorporated to obtain information faster, finding relevant clues even in cases where background information about a crime or misdemeanor is lacking,” Castillo said.
Inauguration of new rooms
The analytics system is part of the Department of Homeland Security’s Protected Neighborhood program. Regarding this initiative, since 2018, a total of 1329 cameras have been installed in different neighborhoods of the municipality, which have Neighborhood Security Centers (NSCs).
In 2020, the integration of analytics began and 43 centers have already been reached, with a total of 886 cameras capable of intelligent detection.
Only in January of this year, three camera projects have already been inaugurated in the CSVs composed of residents of Hendaye, Carlos Alvarado and Medinaceli streets.
How to enter this system?
Residents of Las Condes can manage, through their neighborhood security centers, the installation of these cameras at a reduced cost thanks to municipal co-financing, which can range from 50 to 90 percent of the total, depending on the project.
Regarding specific requests, the following will be considered:
*Directorate of Public Safety camera projects can only be installed on public roads that correspond to residences, buildings or apartments.
For more information about this program and the intelligent detection system, interested residents can write to programasdeseguridad@lascondes.cl.