As an active remote sensing technology, SAR has the ability to work all day and all weather, being hardly affected by bad weather or light conditions. Spaceborne SAR has been widely concerned because of its fast response, wide coverage, and strong anti-strike ability [
1]. With the development of high-resolution spaceborne SAR, the influence of moving target becomes more prominent, which becomes the focus of research [
2,
3]. Moving targets, such as vehicles, trains, ships, and some military targets are the key targets of SAR imaging, which has very important application value in military reconnaissance, ocean observation, and environmental detection. The development of high resolution makes the influence of moving target more prominent, resulting in defocusing and other unexplained phenomena. On the one hand, it affects the interpretation and detection of the target, on the other hand, it also brings the possibility of parameter estimation of the moving target. Generally speaking, it is very important to study the moving target of high-resolution SAR.
Presently, the research on high-resolution SAR moving target mainly focuses on four aspects. The first is the analysis of imaging signatures of moving target. Most of the literatures analyzed the imaging effect of the uniform moving target, and pointed out that the target movement would lead to the target azimuth position dislocation and target defocusing [
4,
5,
6]. There are also literature analyses of the imaging effect of moving targets with acceleration [
7], that is, the third phase error and the asymmetric distortion of the image. Another part of the literature analyzes the influence of the complex motion of the target, such as the rotation motion of the target itself [
8], three-dimensional motion [
9,
10], vibration, and other fretting forms [
11]. There are also a part of literatures that introduces the influence of ship fretting, such as the ship’s three-dimensional swing, including roll, pitch, yaw, and so on [
12,
13]. It is pointed out that the effect of target micro motion becomes stronger with the increase of resolution, mainly due to ghosting, linear defocusing and other phenomena [
14]. The second is ground moving target indication (GMTI) [
15]. GMTI, as a common SAR imaging mode, has been carried on many satellite models, and it has a better performance. GMTI mainly aims at moving target detection and imaging, and the commonly used method is displaced phase center antenna (DPCA) [
16], space-time adaptive processing (STAP) [
17], along-track interferometry (ATI), and so on [
18]. Among all kinds of moving targets, ship is just a kind of research object that attracts much attention. Ship detection is also a very popular research direction in the SAR target detection area. There are two main methods of ship detection. One is the traditional CFAR (Constant False-Alarm Rate) based detection method [
19,
20,
21]. The other is the detection method that is based on deep learning, which has been very popular in recent years [
22,
23,
24]. Of course, there are also attempts to combine the two to get better results [
25]. The third is the estimation of moving target parameters. Most literatures have studied the velocity estimation of uniform moving target. The literature mostly concentrates on the estimation of Doppler parameters dealing with raw data [
26,
27,
28]. For the estimation of the Doppler centroid that was determined by the radial velocity, the nominal-spectrum method, the maximum likelihood method, and the time-domain method are commonly used. For the azimuth velocity estimation, the typical approach uses the Doppler rate filters to select the velocity that better focuses the target or applies matching criteria on the displacement vector that was obtained from the spatial shift of the moving target from sub-apertures [
29,
30]. Some literatures also study the parameter estimation of micro moving target. Canada’s defense research and development department used autocorrelation method to estimate the rotation period of the rotating antenna target [
31]. Subsequently, the Doppler frequency of vibration and rotation targets was analyzed by using smooth pseudo WVD at the University of Zurich, Switzerland. In 2011, the Air Force Academy of Engineering studied the micro Doppler effect of ground rotating parts in Bistatic SAR mode and the method of extracting micro Doppler parameters of ground vibration targets that are based on DPCA clutter suppression [
32]. In recent years, the method of parameter estimation while using sparse reconstruction has also been proposed, and there are related literatures as [
33,
34]. The fourth and last is the focus of moving targets. Focusing imaging for moving targets has always been a hot research direction of scholars. Most of them focus on Doppler parameter estimation compensation imaging based on echo data [
29]. In some literatures, compressed sensing is introduced into the focus of moving targets [
35]. Some papers also study the high-resolution imaging technology of complex moving targets that are based on the imaging principle of inverse synthetic aperture radar (ISAR) [
36,
37].
Presently, the research on the moving target has been comprehensive, but there is no systematic research on the turning motion target. Turning movement is common in trains, high-speed rail, and other targets with a certain length and speed. In recent years, domestic and foreign railways have developed rapidly. In particular, China’s high-speed rail network has been built with the potential of eight horizontal and eight vertical lines, occupying the first place in the world. The scattering intensity of train, high-speed rail is large, and the signatures of its motion in SAR image are more serious with the gradual improvement of SAR resolution. The research on it will benefit the target interpretation and application of SAR image.
This paper focuses on the research of turning motion target. By combining the geometric modeling of turning motion and the imaging geometry of spaceborne SAR, the imaging signatures of turning motion targets are analyzed. The error that is caused by turning movement is quantitatively and qualitatively analyzed. The simulation results and GF-3 images show the correctness of the imaging signatures of the turning motion target analyzed in this paper. Then a method for estimating the turning velocity is proposed. This method uses the phase error compensated in the refocusing process for azimuth velocity calculation and the azimuth offset of target out of track to estimate the radial velocity. The amplitude and direction of the velocity can be obtained by using both of them. The results of simulation and GF-3 data processing verify the correctness of the analysis and method.