Modern drones become more and more intelligent thanks to the integration of in -depth learning and computer vision. In a recent study published in the IPSJ transactions on vision and IT applicationsInnovative methods of detecting and monitoring drones using multi-limited systems and advanced algorithms have been presented.
The researchers proposed a system composed of a high static angle camera and a rotary turret with a high -resolution narrow angle camera. This system can detect small objects at large distances and analyze them in detail using zoomable cameras. At the heart of this technology is a modified YOLOV3 architecture optimized for fast and precise detection.
The system presented in the study consists of several key components, each playing a crucial role in the high precision guarantee and the effectiveness of the detection and monitoring of drones.
Wide angle camera
The wide angle camera is mounted on a stationary platform with a 16 mm focal lens, offering a field of vision of around 110 °. This allows the camera to cover large areas and monitor long distances. The camera can transmit images to a resolution of 2000 x 1700 pixels at a speed of around 25 frames per second. The large field of vision plays a critical role in the initial detection of small objects such as drones on the horizon, allowing the system to quickly respond to new objects in its field of vision.
Rotative turret with narrow angle camera
The second camera of the system is mounted on a rotary turret, allowing it to change its field of vision and follow the objects detected by the wide angle camera. The narrow angle camera is equipped with a 300 mm focal lens, offering a field of vision of around 8.2 ° and the ability to zoom more than 35 times. This camera is designed for a detailed analysis of long distances objects, allowing the system to identify and accurately follow the drones. The turret can turn quickly and adjust the angle of the camera to capture high quality images of target objects.
Main calculation unit based on Linux
The central element of the system is the main calculation unit, which is a computer based on Linux equipped with an NVIDIA graphics processor. This unit deals with the images captured by the cameras and performs deep learning algorithms on the graphic processor, which is the Nvidia GeForce K620 with 2 GB of memory. The use of the GPU allows the system to process large volumes of data in real time and ensures high performance in the execution of complex calculation tasks. Deep learning algorithms, such as Yolov3, have been modified and optimized to work in this system, reaching high precision and detection speed.
These components work in close integration to ensure high reliability and efficiency of the drone detection and monitoring system. The interaction between the wide angle and narrow angle cameras, as well as the powerful calculation unit, allows the system to respond quickly and precisely to the appearance of drones in its field of vision, guaranteeing a high level of safety and control.
For drone detection, the system uses a modified version of Yolov3, which effectively treats images and detects small objects. Unlike standard approaches, this method uses a regression model for the rapid location of objects in the images. The modification of the YOLOV3 architecture included the reduction in the number of filters while retaining the number of layers, optimizing the operation of the system on limited GPU resources.
This innovative system guarantees high precision and detection speed, which makes it ideal for use in safety and surveillance tasks. He is able to detect drones on the horizon, to follow their movements and, if necessary, to analyze them using the camera with narrow angle.
Research demonstrates the potential of in -depth learning in improving the capacities of drones to perform complex time tasks, including surveillance, safety and rescue operations.
