The Argus robot is designed to detect people, cars, and other objects by using a smart thermal imaging surveillance system. The thermal imaging dual-spectrum camera that is placed on the robot is capable of working under the cover of darkness, and in conditions where there is condensate from fog and rain, with no external illumination. The combination of the robot’s mobility and long-range surveillance capability makes using Argus ideal for addressing security related challenges in critical areas on guarded premises.
The robot is equipped with an advanced intelligence video analytics system that allows detecting and tracking objects. Using deep learning algorithms permits reliably detecting a person at distances up to 200 meters in complete darkness. The movement of the robot as part of a group is controlled by risk assessment algorithms with elements of artificial intelligence. The Argus security robot allows substantially boosting the capabilities of a robot patrol group.
Along with the dual-spectrum PTZ camera, the robot has 6 high-resolution cameras that provide continuous 360-degree surveillance in the immediate vicinity of the robot. These cameras allow controlling the space surrounding the robot, regardless of where the PTZ camera is directed at any given moment. Panoramic cameras that work in the visible spectrum have high sensitivity, but not enough to get a high-quality image in complete darkness.
Two robot models are produced that are distinguished by the presence of an on-board system for detecting people and human faces from the images received from panoramic cameras. The S5.2 IR Argus robot does not have that function, but is capable of working for a longer period of time without being charged. The set of equipment it has is ideal for work during nighttime on premises that do not have any artificial illumination. The S 5.2 IR IS Hershel robot is fitted out with an additional embedded computer to detect people and isolate their faces. People are recognized in close proximity to the panoramic cameras, either with artificial lighting at night or with natural light in the daytime. The detection range for a person is up to 30 meters in the daytime, and the range for obtaining a facial photo that is suitable for further recognition is up to 3 meters. Both robot models are able to detect and track objects using a thermal PTZ camera under any lighting conditions.
The Argus robot is ideal for perimeter and fence line security. The robot is optimal for remote monitoring from positions that are inaccessible to potential intruders. This model has the option of installing a mesh fence detector to avoid colliding with a less visible obstacle.
The Hershel robot was created to patrol premises at manned and unmanned facilities. At the former, it is possible to recognize personnel by their faces, while at the latter detecting the presence of a person near the robot is categorized as an unconditional alarm situation.
Using thermal robots to guard critical facilities puts forth stringent requirements for the quality of the robot’s on-board video analytics system. To avoid false alarms, software solutions are used with the application of deep learning algorithms based on neural networks.
The thermal dual-spectrum PTZ camera scans the surrounding space by rotating in a circular manner. The video data from both PTZ camera sensors is analyzed by a supercomputer that uses an NVIDIA processor. Despite the high sensitivity and adequate resolution possessed by the passive thermal camera, analyzing images received from the camera in the visible spectrum during the daytime allows obtaining additional information that is useful to detect an object without making any errors.
When a given object is detected in the camera’s field of view (this could be people, cars, tracks, bicycles, horses, or camels), the embedded computer puts the camera in tracking mode for the object detected and transfers the image to the video management system control room. Thanks to the phenomenal processing power of the onboard computer, every frame of the video is subjected to deep-learning analysis by the dual-spectrum camera, ensuring high reliability for tracking of detected objects and reducing false alarms.
When patrolling in conditions where there is sufficient illumination, it is advisable to use intelligent robots equipped with high-resolution visible-spectrum cameras. The S5.2 IS Promt robot is designed to patrol pedestrian passageways and recognize people in the robot’s immediate vicinity. The AI S5.2 PTZ IS Picard robot was created to detect people and perform smart video surveillance at distances up to 100 meters.
Skillful use of the features inherent in each robot model allows fitting out even the most complex facilities with robotic security. The unique AI capabilities inherent in the management of patrol routes and determining the positions used for surveillance provide a very high level of security.
Thermal dual-spectrum robots, like robots of other models, are equipped with powerful built-in artificial intelligence. This makes robotic security extremely reliable. Each robot autonomously makes decisions, depending on the location of other robots and the results of big data analysis of risk assessment. The failure of one or several robots, as well as the absence of a single management server, does not lead to a fatal loss of security for the area under guard.