Argustec solves three major problems in forest fire alarm

01. Fewer false alarms

In the forest fire monitoring environment, there are water mist, clouds and other objects similar to smoke, as well as high-temperature objects such as vehicle engines and water surface reflections, which will greatly interfere with the normal recognition of the fireworks algorithm and cause many false alarms. The workload of false alarm verification for operation and maintenance personnel is large, and it is difficult to detect the correct alarm in time.

Argustec has conducted in-depth research on the characteristics of fireworks, carried out special research and optimization on the sources of false alarms, and innovatively developed a multi-algorithm engine solution based on deep learning. It has built-in multiple fireworks false alarm filtering algorithms, which can intelligently filter common false alarms such as clouds, chimneys, and vehicle engines. The current project has reduced the false alarm rate of fire points to 0.25 units/day, and the false alarm rate of fireworks to 0.16 units/day.

02. Accurate positioning

After a fire is discovered, it is necessary to accurately locate the fire point and quickly put out the fire and rescue. However, due to the influence of altitude, equipment geographic coordinates, equipment horizontal reference azimuth, equipment vertical pitch angle and DEM map, the fire point positioning error presented by traditional PTZ on the map is generally more than 1km, which greatly affects the timeliness of firefighters locating the fire point and putting out the fire.

Argustec has added a unique attitude perception module to the equipment, which can assist the equipment in intelligently sensing the angular deviation from the sea level and achieving consistency with the GIS map benchmark zero point. At the same time, it has innovatively developed a spatial perception module that can associate the mechanical structure of the fire prevention camera with the spatial coordinate system, greatly improving the perception accuracy of the camera's intelligent recognition system for spatial position, and effectively improving the accuracy of fire point positioning. In a certain project, the actual measured fire point positioning accuracy at a distance of 4km can reach 63m.

03. Stable operation

In forest fire prevention applications, the equipment is usually installed on a high tower in a remote area, which makes operation and maintenance extremely troublesome. Therefore, the stability of the equipment and the guarantee of localized operation and maintenance services are very important.

In order to solve the common problems in outdoor operation, such as fogging leading to blurred lenses, dirty lenses due to long-term operation, disconnection caused by unstable power supply, and the service life of continuously zooming and focusing lenses, Argustec adopted different technical solutions in structural design to solve them: