1. Shool of Military Biomedical Engineering, Air Force Medical University, Xi’an Shaanxi 710032, China丨
2. Drug and Instrument Supervisory & Test Station, Xining Joint Logistics Support Center, Lanzhou Gansu 730050, China丨
3. Department of Radiology, Tangdu Hospital, Air Force Medical University, Xi’an Shaanxi 710038, China
Abstract:Objective To improve the rate of the wounded identification and solve the problem of finding the wounded in similar
colour clothes quickly under low-contrast background environment, a new identification scheme based on hyperspectral of uncrewed
aerial vehicle is proposed. Methods The remote sensing images of the camouflaged wounded on the ground were collected by the
M600 six rotor UAV equipped with HY-1010 hyperspectral camera, and the simulated blood spraying camouflage clothes were used
to simulate the wounded under low-contrast background. After the hyperspectral remote sensing images were preprocessed, the target
spectrum and background spectrum were extracted, and the three common hyperspectral ground object recognition methods of mixed
tuned matched filtering, constrained energy minimization (CEM) and adaptive coherence estimator (ACE) were used to identify the
wounded. Results The three recognition methods can identify six simulated wounded. From the statistical analysis of pixel level, the
false recognition rate of ACE method was the lowest 53.2%, and the correct recognition rate of CEM method was the highest 50%.
Conclusion This experiment uses hyperspectral recognition technology to identify the bloodstain on the camouflage clothes in the
grassland environment, which provides a new idea for the identification of the wounded under low-contrast background environment
朱明明1,2,尚丹婷3,曹育森1,雷涛1,夏娟娟1,李钊1,景裕1,张林媛1,王健琪1,路国华1. 基于高光谱的低对比度伤员目标搜寻技术研究[J]. 中国医疗设备, 2021, 36(6): 5-10.
ZHU Mingming1,2,Shang Danting3,CAO Yusen1,LEI Tao1,XIA Juanjuan1,LI Zhao1,JING Yu1,ZHANG Linyuan1,WANG Jianqi1,LU Guohua1. Research on Low-Contrast Wounded Human Target Search Technology Based on Hyperspectral. China Medical Devices, 2021, 36(6): 5-10.