Zhang Bo, Wang Chao, Zhang Hong, Wu Fan, Tang Yi-Xian
Key Laboratory of Digital Earth Science, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, China.
Sensors (Basel). 2017 Feb 6;17(2):298. doi: 10.3390/s17020298.
By acquiring information over a wide area regardless of weather conditions and solar illumination, space-borne Synthetic Aperture Radar (SAR) has the potential to be a promising application for traffic monitoring. However, the backscatter character of a vehicle in a SAR image is unstable and varies with image parameters, such as aspect and incidence angle. To investigate vehicle detectability in SAR images for traffic monitoring applications, images of four common types of vehicles in China were acquired using the fully polarimetric (FP) SAR of Radarsat-2 in our experiments. Methods for measuring a vehicle's aspect angle and backscatter intensity are introduced. The experimental FP SAR images are used to analyze the detectability, which is affected by factors such as vehicle size, vehicle shape, and aspect angle. Moreover, a new metric to improve vehicle detectability in FP SAR images is proposed and compared with the well-known intensity metric. The experimental results show that shape is a crucial factor in affecting the backscatter intensity of vehicles, which also oscillates with varying aspect angle. If the size of a vehicle is smaller than the SAR image resolution, using the intensity metric would result in low detectability. However, it could be improved in an FP SAR image by using the proposed metric. Compared with the intensity metric, the overall detectability is improved from 72% to 90% in our experiments. Therefore, this study indicates that FP SAR images have the ability to detect stationary vehicles on the road and are meaningful for traffic monitoring.
通过在不受天气条件和太阳光照影响的广阔区域获取信息，星载合成孔径雷达（SAR）有潜力成为交通监测的一项有前景的应用。然而，SAR图像中车辆的后向散射特性不稳定，会随图像参数（如方位角和入射角）而变化。为了研究用于交通监测应用的SAR图像中车辆的可检测性，我们在实验中使用Radarsat - 2的全极化（FP）SAR获取了中国四种常见类型车辆的图像。介绍了测量车辆方位角和后向散射强度的方法。利用实验性的FP SAR图像分析可检测性，其受到车辆尺寸、车辆形状和方位角等因素的影响。此外，提出了一种提高FP SAR图像中车辆可检测性的新指标，并与著名的强度指标进行比较。实验结果表明，形状是影响车辆后向散射强度的关键因素，后向散射强度也会随着方位角的变化而振荡。如果车辆尺寸小于SAR图像分辨率，使用强度指标会导致可检测性较低。然而，通过使用所提出的指标，在FP SAR图像中可检测性可以得到提高。在我们的实验中，与强度指标相比，整体可检测性从72%提高到了90%。因此，本研究表明FP SAR图像有能力检测道路上的静止车辆，对交通监测具有重要意义。
