Objective  In recent years, with the gradual opening of low altitude airspace and the continuous development of unmanned aerial vehicles (UAVs) technology, rapidly developing drones have been widely used in military, agriculture, transportation, public safety and other fields. However, due to the characteristics of simple operation, low cost, difficulty in supervision, and strong breakthrough ability, UAVs have caused many safety accidents and violent threats to social security and stability. Therefore, there is a strong demand for countermeasures against UAVs. With the development of laser weapons, the technology is becoming increasingly mature and has huge advantages in anti-UAVs. Under the action of high-energy laser, the target is prone to burning and catching fire, resulting in the target being blocked by flame smoke. Currently, visible light and infrared detection methods are easily affected by severe interference from flame and smoke, resulting in poor or unclear imaging of targets. Therefore, aiming at the problem that visible light and infrared detection methods cannot achieve stable tracking and aiming under flame and smoke, the paper studies a high-precision aiming scheme based on active lidar system.

  Methods  The paper proposes a high-precision aiming scheme based on the active lidar system. Firstly, the laser characteristics under the flame and smoke were analyzed. Secondly, a lidar detection system based on APD single photon detector was designed. Then, the theoretical curve of the detection probability was simulated and analyzed as a function of the laser pulse energy. Finally, a photon imaging system based on InGaAs-SPAD was built, and the imaging system was tested and experimentally conducted indoors.

  Results and Discussions  In the absence of the influence of fire and smoke, the X and Y coordinate trajectory curves of the target centroid based on visible light image tracking and range image tracking are basically consistent. By calculating the angular radian size of the target centroid position based on distance image tracking relative to the visible light tracking target centroid position, the X-coordinate angular radian curve and Y-coordinate angular radian curve were obtained (Fig.8). It is known that the average angular radian of the X-coordinate is about 0.55 mrad, and the average angular radian of the Y-coordinate is 0.53 mrad. By using a lidar system to collect range profiles under smoke, when the gating delay is 0 ns, the target is obstructed by smoke on the collected range profiles, making it impossible to obtain the image of the UAV (Fig.9). This is because when photons scattered by smoke cause detector response, photons reflected by the target will not cause detection starting point response, thus unable to obtain the range profiles of the target. Subsequently, through debugging under the same conditions, we obtained the range profiles for filtering out smoke obstruction when setting the gating delay to 43 ns (Fig.10). So active laser imaging can filter out the influence of smoke. In the experiment, the UAV was exposed to strong light during its movement, and under the sudden change conditions, the target lost track (Fig.12). Due to the correlation between the loss of target tracking and the impact of the blue part of the range profile mutation, when tracking based on the range profile, the influence of firelight noise on the target is filtered out, and the target is tracked (Fig.13). The deviation of the X and Y coordinates of the tracking rectangle center from the target centroid is 0.58 mrad and 0.39 mrad, and the target can be stably tracked within the range of the tracking rectangle under sudden changes in firelight.

  Conclusions  The analysis of the optical characteristics of the firelight background and the laser attenuation characteristics of smoke shows that the radiance at 1 064 nm wavelength within the 10 nm narrowband filter passband is approximately 0.014 W/cm²; The transmittance range of laser in smoke is 60%-85%; In the case of backward scattering, f_r is approximately 0.2-8 in the rough range of 0.2-0.8 μm. A lidar detection system is designed based on APD single photon detector, simulation analysis of the theoretical curve of detection probability changing with laser pulse energy. Under the conditions of this experiment, the optimal range for obtaining laser pulse energy is 0.75-4 mJ. Based on the above analysis, experimental verification of UAV lidar detection was conducted under smoke and fire backgrounds. Under the background without the influence of fire and smoke, the relative offset of the target centroid position based on range profile tracking to the target centroid position based on visible light image tracking is less than 0.55 mrad. Under simulated smoke conditions, clear contour target distance profiles were obtained by filtering out smoke effects through gating delay; Under simulated firelight mutation conditions, the X and Y coordinates of the tracking box center deviate from the target centroid by 0.58 mrad and 0.39 mrad. The experimental results show that this scheme can achieve imaging of unmanned aerial vehicles in the background of flame and smoke, and the Mean-Shift algorithm is used for range profile tracking, which compensates for the shortcomings of lighting mutation and easy loss of tracking when the background is similar to the target based on visible light image sequence tracking.