Automatic Ship Detection in Remote Sensing Images from Google Earth of Complex Scenes Based on Multiscale Rotation Dense Feature Pyramid Networks

By Xue Yang, Kun Fu, Jirui Yang, Xian Sun, Menglong Yan, Zhi Guo. Remote Sensing, 2018

505 citations recorded by the source corpus.

Ship detection has been playing a significant role in the field of remote sensing for a long time, but it is still full of challenges. The main limitations of traditional ship detection methods usually lie in the complexity of application scenarios, the difficulty of intensive object detection, and the redundancy of the detection region. In order to solve these problems above, we propose a framework called Rotation Dense Feature Pyramid Networks (R-DFPN) which can effectively detect ships in different scenes including ocean and port. Specifically, we put forward the Dense Feature Pyramid Network (DFPN), which is aimed at solving problems resulting from the narrow width of the ship. Compared with previous multiscale detectors such as Feature Pyramid Network (FPN), DFPN builds high-level semantic feature-maps for all scales by means of dense connections, through which feature propagation is enhanced and feature reuse is encouraged. Additionally, in the case of ship rotation and dense arrangement, we design a rotation anchor strategy to predict the minimum circumscribed rectangle of the object so as to reduce the redundant detection region and improve the recall. Furthermore, we also propose multiscale region of interest (ROI) Align for the purpose of maintaining the completeness of the semantic and spatial information. Experiments based on remote sensing images from Google Earth for ship detection show that our detection method based on R-DFPN representation has state-of-the-art performance.

Subfields: Archeology, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/dad0e3d7b1dc51196255ba220ed58ff29aa7c92b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Archeology

1. Xue Yang
2. Junchi Yan
3. Wentao Wang
4. Qi Ming
5. Yue Zhou
6. Wentong Li
7. Jianke Zhu
8. Liping Hou

In Computer Vision and Pattern Recognition

1. Junchi Yan
2. Xue Yang
3. Guisong Xia
4. Wentao Wang
5. Jian Ding
6. Jianke Zhu
7. Wentong Li
8. Qi Tian

Citations in the corpus

• Towards Multi-class Object Detection in Unconstrained Remote Sensing Imagery (2018, background): “Another alternative is image or feature pyramids [7, 12, 14, 20, 21, 31].”
• SCRDet++: Detecting Small, Cluttered and Rotated Objects via Instance-Level Feature Denoising and Rotation Loss Smoothing (2020, background): “6a shows the rectangular definition of the 90 angle representation range [18], [30], [47], [58], [59].”
• Oriented RepPoints for Aerial Object Detection (2021, methodology): “Object detection emerges as an effective tool for aerial images analysis [1], [2], [11], [23].”
• Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark (2022, unknown): “, 2018) and Deep Feature Pyramid Network (DFPN) (Yang et al., 2018) are presented.”
• Learning RoI Transformer for Oriented Object Detection in Aerial Images (2019, methodology): “The [30, 43, 41, 1] used a design of rotated anchor in RPN [32].”
• Dynamic Anchor Learning for Arbitrary-Oriented Object Detection (2020, result): “It can be seen that our method further Baseline (Yang et al. 2018) HAMBox(Liu et al. 2020 improves the performance by 2.83%, reaches the mAP of 76.95%, achieving the best results among all compared models.”
• RBox-CNN: rotated bounding box based CNN for ship detection in remote sensing image (2018, methodology): “Recently, similar to RRPN and R 2 CNN, Yang et al. [11] proposed a detection framework called rotation dense feature pyramid networks (R-DFPN) based on Faster R-CNN for ship detection.”
• Learning Modulated Loss for Rotated Object Detection (2019, methodology): “The existing region-based rotated object detectors usually regress five parameters (coordinates of the central point, width, height, rotation angle) [10, 19, 20, 21] to describe rotated bounding boxes and use (cid:96) 1 -loss as loss functions.”
• Learning High-Precision Bounding Box for Rotated Object Detection via Kullback-Leibler Divergence (2021, methodology): “Extending the above horizontal case, existing rotation detection models [1, 12, 13, 14, 15] also use regression loss which simply involves an extra angle parameter θ: tpθ = f(θp − θa), t t θ = f(θt − θa) (3) where f(·) is used to deal with angular periodicity, such as trigonometric functions, modulo”
• Dense Label Encoding for Boundary Discontinuity Free Rotation Detection (2020, methodology): “Many rotation detectors (including those for aerial image [46, 15, 47, 2, 6, 51, 45, 30, 48, 49, 50], scene text [56, 23, 14, 26, 19, 18] and face [34, 13, 33]) are derived based on the vanilla detection algorithms [10, 32, 31, 5, 20, 21].”
• CAD-Net: A Context-Aware Detection Network for Objects in Remote Sensing Imagery (2019, methodology): “For example, [29] uses multi-scale CNN features for airport detection, [31] takes a weakly supervised learning framework for aircraft detection using CNNs, [30] and [32] integrate dense FPN (DFPN) and rotation regression for ship detection in remote sensing images, and [28] uses CNN features for mul”
• Dynamic Refinement Network for Oriented and Densely Packed Object Detection (2020, unknown): “20 R-DFPN [41] 80.”
• R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object (2019, background): “Experiments show that FRM has greatly improved the category that are not fully learned due to the small number of samples and inaccurate features, such as BD, GTF, BC, SBF, RA, HC, which increased by 4.09%, 2.83%, 3.4%, 4.82%, 1.22%, and 19.26%, respectively.”