Hongjun Song

Hongjun Song


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 January 2020 in Remote Sensing
Verified
Journal Article
Open access
PDF file
JZJiahuan Zhang
Jiahuan Zhang
HSHongjun Song
Hongjun Song
BZBinbin Zhou
Binbin Zhou
Synthetic aperture radar (SAR) has become one of the most important means of information acquisition in today’s society and shows great potential in many fields. Target identification and classification of SAR images are also the focus of research. With the vigorous development of deep learning, many researchers apply this method to SAR target classification to obtain a more automatic process and more accurate results. In this paper, a novel deep forest model constructed by multi-grained cascade forest (gcForest), which is different from the traditional neural network (NN) model, is employed to classify ten types of SAR targets in the moving and stationary target acquisition and recognition (MSTAR) dataset. Considering that the targets of input images may be off-center and of different sizes in practical applications, two improved models based on varying weights by image features have been put forward, and both obtain better results. A series of experiments have been conducted to optimize model parameters, and final results with the MSTAR dataset illustrate that the two improved models are both superior to the original gcForest model. This is the first attempt to classify SAR targets using the non-NN model.

12 July 2021 in Remote Sensing
Verified
Journal Article
Open access
PDF file
XJXinlin Jia
Xinlin Jia
HSHongjun Song
Hongjun Song
WHWenjing He
Wenjing He
As an active microwave remote sensing device, synthetic aperture radar (SAR) has been widely used in the field of marine surveillance. However, moving ships appear defocused in SAR images, which seriously affects the classification and identification of ships. Considering the three-dimensional (3-D) rotational motions (roll, pitch, and yaw) of the navigating ship, a novel method for refocusing moving ships in SAR images based on inverse synthetic aperture radar (ISAR) technique is proposed. First, a rectangular window is used to extract the defocused ship subimage. Next, the subimage is transformed into the ISAR equivalent echo domain, and the range migration and phase error caused by the identical movement of all ship scatterers are compensated. Then, the optimal imaging time can be selected by the maximum image contrast search method. Finally, the iterative adaptive approach (IAA) is used to obtain the image with high resolution. This method has satisfactory imaging performance in both azimuth resolution and image focus, and the amount of calculation is small due to the processing of subimages. Simulated data and Gaofen-3 real SAR data are used to verify the effectiveness of the proposed method.

11 January 2020 in Sensors
Verified
Journal Article
Open access
PDF file
JLJin Liu
Jin Liu
WWWei Wang
Wei Wang
HSHongjun Song
Hongjun Song
Weighting window functions are commonly used in Synthetic Aperture Radar (SAR) imaging to suppress the high Peak SideLobe Ratio (PSLR) at the price of probable Signal-to-Noise Ratio (SNR) loss and mainlobe widening. In this paper, based on the method of designing a mismatched filter, we have proposed a Quadratically Constrained Quadratic Program (QCQP) approach, which is a convex that can be solved efficiently, to optimize the weighting window function with both amplitude and phase, expecting to offer better imaging performance, especially on PSLR, SNR loss, and mainlobe width. According to this approach and its modified form, we are able to design window functions to optimize the PSLR or the SNR loss under different kinds of flexible and practical constraints. Compared to the ordinary real-valued and symmetric window functions, like the Taylor window, the designed window functions are complex-valued and can be asymmetric. By using Synthetic Aperture Radar (SAR) point target imaging simulation, we show that the optimized weighting window function can clearly show the weak target hidden in the sidelobes of the strong target.

23 February 2021 in Remote Sensing
Verified
Journal Article
Open access
PDF file
JZJiahuan Zhang
Jiahuan Zhang
HSHongjun Song
Hongjun Song
Target detection on the sea-surface has always been a high-profile problem, and the detection of weak targets is one of the most difficult problems and the key issue under this problem. Traditional techniques, such as imaging, cannot effectively detect these types of targets, so researchers choose to start by mining the characteristics of the received echoes and other aspects for target detection. This paper proposes a false alarm rate (FAR) controllable deep forest model based on six-dimensional feature space for efficient and accurate detection of weak targets on the sea-surface. This is the first attempt at the deep forest model in this field. The validity of the model was verified on IPIX data, and the detection probability was compared with other proposed methods. Under the same FAR condition, the average detection accuracy rate of the proposed method could reach over 99.19%, which is 9.96% better than the results of the current most advanced method (K-NN FAR-controlled Detector). Experimental results show that multi-feature fusion and the use of a suitable detection framework have a positive effect on the detection of weak targets on the sea-surface.

12 April 2021 in Remote Sensing
Verified
Journal Article
Open access
PDF file
WHWenjing He
Wenjing He
HSHongjun Song
Hongjun Song
YYYuanyuan Yao
Yuanyuan Yao
XJXinlin Jia
Xinlin Jia
Road network is an important part of modern transportation. For the demands of accurate road information in practical applications such as urban planning and disaster assessment, we propose a multiscale method to extract road network from high-resolution synthetic aperture radar (SAR) images, which consists of three stages: potential road area segmentation, preliminary network generation, and road network refinement. Multiscale analysis is implemented using an image pyramid framework together with a fixed-size filter. First, a directional road detector is designed to highlight road targets in feature response maps. Subsequently, adaptive fusion is performed independently at each image scale, followed by a threshold method to produce potential road maps. Then, binary maps are decomposed according to the obtained direction information. For each connected component (CC), quality evaluation is conducted to further distinguish road segments and polynomial curve fitting is adopted as a thinning method. Multiscale information fusion is realized through the weighted sum of road curves. Finally, tensor voting and spatial regularization are employed to generate the final road network. Experiments on three TerraSAR images demonstrate the effectiveness of the proposed algorithm to extract road network completely and correctly.