To address the strong reliance of existing monocular speed estimation methods on key prior information such as target size, vanishing point position, and scene markers, a new method for estimating the speed of multiple ground military targets from a drone perspective is proposed. Firstly, the trajectory information of moving targets in consecutive frame sequences is extracted through multi-target tracking technology, and the time interval is calculated in combination with the video frame rate. Secondly, a geometric projection model is constructed using the flight height of the drone and the camera gimbal attitude angle to establish the mapping relationship between the pixel coordinate system and the world coordinate system, and then the real displacement distance of the moving target is calculated. Finally, the performance of the proposed speed measurement algorithm is verified through experiments, and the calculation accuracy of the geometric projection model is quantitatively analyzed and evaluated. The experimental results show that when the flight height is greater than 20.0 m and the attitude angle is less than 30.0°, the relative measurement error does not exceed 2%. Compared with other speed measurement methods, the proposed method meets the requirements of speed estimation for moving targets in battlefield situation awareness without relying on prior information such as target size, vanishing point position, and scene markers, providing a more adaptable new method for speed estimation.
XUTianxu, MOXilong, YUWei. On the potential of UAVs in future military operations[J]. Defence Industry Conversion in China, 2025(2): 37-38. (in Chinese)
ZHANGXiaoling, LUOYinsheng, ZHANGBaofeng. Measurement on velocity of dynamic object in long distance based on binocular vision system[J]. Journal of Optoelectronics·Laser, 2016, 27(9): 935-940. (in Chinese)
CHANGZiting, SHIYuqing, WANGJun, et al. Vehicle speed measurement method based on binocular vision[J]. Computer Science, 2021, 48(9): 135-139. (in Chinese)
[12]
鱼朋波. 高速公路机器视觉测速方法研究[D]. 沈阳: 沈阳航空航天大学, 2014.
[13]
KER, KIMS, LIZ, et al. Motion-vector clustering for traffic speed detection from UAV video[C]//2015 IEEE First International Smart Cities Conference (ISC2), 2015: 1-5.
[14]
LONGH, CHUNGY N, LIJ D. Automatic vehicle speed estimation method for unmanned aerial vehicle images[J]. Journal of Information Hiding and Multimedia Signal Processing, 2018, 9(2): 442-451.
LIANGJianshu, JIAHuilun, ZHANGLinlin. Research on algorithm of traffic accident vehicle speed based on monitoring video image[J]. Agricultural Equipment & Vehicle Engineering, 2022, 60(6): 87-89. (in Chinese)
[17]
BISWASD, SUH, WANGC, et al. Speed estimation of multiple moving objects from a moving UAV platform[J]. ISPRS International Journal of Geo-Information, 2019, 8(6): 259.
CHENBin, LIUHuazhou, LIBocheng, et al. Monocular distance measurement method for insulator in unmanned aerial vehicle inspection images based on YOLOv5 algorithm[J]. Radio Engineering, 2024, 54(6): 1421-1430. (in Chinese)
XINGShuhao, DENGJie, HUZhe, et al. A method for measuring ship’s sailing speed based on uav aerial images[J]. Digital Communication World, 2023(3): 100-102. (in Chinese)
[22]
LIJ, CHENS, ZHANGF, et al. An adaptive framework for multi-vehicle ground speed estimation in airborne videos[J]. Remote Sensing, 2019, 11(10): 1241.
[23]
SILVA BASTOS M EDA, FREITASV Y F, DE MENEZESR S T, et al. Vehicle speed detection and safety distance estimation using aerial images of Brazilian highways[C]//Anais do Seminário Integrado de Software e Hardware (SEMISH 2020), 2020: 258-268.
ZHAOJialiang, LIShizhong, ZHAOZiliang. UAV attitude and heading algorithm based on extended Kalman filter[J]. Journal of North University of China(Natural Science Edition), 2023, 44(3): 305-310.(in Chinese)
ZHENGKai, ZHENGXianmin, YINShaofeng, et al. A UAV velocity measurement approach for moving targets based on least squares fitting algorithm[J]. Journal of Ordnance Equipment Engineering, 2021, 42(3): 61-64. (in Chinese)