结合视觉显著性和EfficientNetV2的舰船目标检测方法

doi:10.3778/j.issn.1002-8331.2210-0267

摘要/Abstract

摘要： 随着光学遥感图像分辨率逐渐提高，对海面舰船目标快速精准检测成为海事研究的基本挑战之一。为了解决检测过程中面临的待检测图像尺寸大而目标稀疏、复杂环境干扰、目标提取时效性差、模型体积计算量大等问题，提出一种实用的舰船检测方案。引入视觉显著性有效加速预筛选过程，利用小波分解系数表达舰船目标区域与背景的差异，抑制噪声的同时增强目标方向特征，通过改进的四元数傅里叶变换相位谱模型（phase spectrum of quaternion Fourier transform，PQFT）生成显著图，并采用Gini指数引导多尺度显著图融合以增强图像尺度适应性及小目标显著性。与其他显著性方法相比，提出的模型能够有效抑制云、雾、海杂波、舰船尾迹等复杂环境的干扰，与经典的滑动窗口或其他区域建议方法相比产生更小的候选区域集合。得到显著图映射后，采用自适应阈值OTSU法对显著图进行二值分割。在目标判别阶段，利用轻量化网络EfficientNetV2有效剔除虚警。实验结果表明，所提出的船舶检测方法鲁棒性高，准确率高达96%，满足实时性需求。

关键词: 光学遥感, 舰船检测, PQFT算法, 视觉显著性, EfficientNetV2

Abstract: With the increasing resolution of optical remote sensing images, fast and accurate detection of ship targets on the sea has become one of the basic challenges of maritime research. In order to solve the problems faced in the detection process, such as large image size but sparse targets, complex background interference, poor timeliness of target extraction, and large calculation of model volume, a practical ship detection scheme is proposed. Visual saliency is introduced to effectively accelerate the pre-screening process, and the difference between the ship target area and the background is effectively expressed by wavelet decomposition coefficients, which can enhance the target directional characteristics while suppressing noise. Saliency map is generated through the improved model based on phase spectrum of quaternion Fourier transform (PQFT). In addition, Gini index is exploited to guide multi-scale saliency image fusion to enhance image scale adaptability and small target saliency. Comparing with other saliency methods, the proposes model can effectively suppress the interference of complex environments such as cloud, fog, sea clutter, and ship wake. More importantly, it produces a smaller set of candidate regions than the classical sliding window or other region recommendation methods. After the saliency map is obtained, the adaptive threshold OTSU method is employed for binary segmentation of saliency map. In the target discrimination stage, the lightweight network EfficientNetV2 is exploited to effectively eliminate false alarms. The experimental results show that the proposes ship detection method has high robustness and accuracy up to 96%, meeting the real-time requirements.

Key words: optical remote sensing, ship detection, phase spectrum of quaternion Fourier transform (PQFT) algorithm, visual saliency, EfficientNetV2

梁秀雅, 冯水春, 陈红珍. 结合视觉显著性和EfficientNetV2的舰船目标检测方法[J]. 计算机工程与应用, 2024, 60(5): 259-270.

LIANG Xiuya, FENG Shuichun, CHEN Hongzhen. Ship Target Detection Method Combining Visual Saliency and EfficientNetV2[J]. Computer Engineering and Applications, 2024, 60(5): 259-270.

参考文献

[1] SCHWEGMANN C P, KLEYNHANS W, SALMON B P. Synthetic aperture radar ship detection using Haar-like features[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14: 154-158.
[2] SHUAI T, SUN K, WU X N, et al. A ship target automatic detection method for high-resolution remote sensing[C]//IEEE International Geoscience and Remote Sensing Symposium (IGARSS), 2016: 1258-1261.
[3] LENG X G, JI K F, XING X W, et al. Area ratio invariant feature group for ship detection in SAR imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(7): 2376-2388.
[4] ZHOU H T, ZHUANG Y, CHEN L, et al. Signal and information processing, networking and computers[M]. 3rd ed. Springer: Singapore, 2018: 164-171.
[5] BORJI A, CHENG M M, JIANG H, et al. Salient object detection: a benchmark[J]. IEEE Transactions on Image Processing, 2015, 24: 5706-5722.
[6] YANG F, XU Q Z, LI B, et al. Ship detection from thermal remote sensing imagery through region-based deep forest[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15: 449-453.
[7] DONG C, LIU J H, XU F, et al. Ship detection from optical remote sensing images using multi-scale analysis and Fourier HOG descriptor[J]. Remote Sensing, 2019, 11(13): 1529.
[8] PANG J M, LI C, SHI J P, et al. R2-CNN: fast tiny object detection in large-scale remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(8): 5512-5524.
[9] 史文旭, 江金洪, 鲍胜利. 基于特征融合的遥感图像舰船目标检测方法[J]. 光子学报, 2020, 49(7): 57-67.
SHI W X, JIANG J H, BAO S L. Ship detection method in remote sensing image based on feature fusion[J]. Acta Photonica Sinica, 2020, 49(7): 57-67.
[10] 陈冬, 句彦伟. 基于改进型YOLOv3的SAR图像舰船目标检测[J]. 系统工程与电子技术, 2021, 43(4): 937-943.
CHEN D, JU Y W. Ship detection in SAR image based on improved YOLOV3[J]. Systems Engineering and Electronics, 2021, 43(4): 937-943.
[11] CHEN J J, XIE F Y, LU Y Y, et al. Finding arbitrary-oriented ships from remote sensing images using corner detection[J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17(10): 1712-1716.
[12] 朱炳宇, 刘朕, 张景祥. 融合Grad-CAM和卷积神经网络的COVID-19检测算法[J]. 计算机科学与探索, 2022, 16(9): 2108-2120.
ZHU B Y, LIU L, ZHANG J X. COVID-19 detection algorithm combining Grad-CAM and convolutional neural network[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(9): 2108-2120.
[13] 王自全, 张永生, 于英, 等. 深度学习背景下视觉显著性物体检测综述[J]. 中国图象图形学报, 2022, 27(7): 2112-2128.
WANG Z Q, ZHANG Y S, YU Y, et al. Review of deep learning based salient object detection[J]. Journal of Image and Graphics, 2022, 27(7): 2112-2128.
[14] SONG Z, SUI H, HUA L. A hierarchical object detection method in large-scale optical remote sensing satellite imagery using saliency detection and CNN[J]. International Journal of Remote Sensing, 2021, 42(8): 2827-2847.
[15] 丁荣莉, 李杰, 沈霁, 等. 基于显著性的遥感图像舰船标检测[J]. 计算机仿真, 2021, 38(11): 5-8.
DING R L, LI J, SHEN J, et al. Ship target detection based on saliency in remote sensing image[J]. Computer Simulation, 2021, 38(11): 5-8.
[16] NIE T, HAN X, HE B, et al. Ship detection in panchromatic optical remote sensing images based on visual saliency and multi-dimensional feature description[J]. Remote Sensing, 2020, 12(1): 152.
[17] 黎经元, 厉小润, 赵辽英. 融合空频域特征的光学遥感图像舰船目标检测[J]. 激光与光电子学进展, 2021, 58(4): 357-365.
LI J Y, LI X R, ZHAO L Y. Ship target detection in optical remote sensing images based on spatial and frequency features[J]. Laser and Optoelectronics Progress, 2021, 58(4): 357-365.
[18] 张丽杰, 尹莉媛, 贾一新. 基于PQFT模型的遥感图像舰船目标检测方法[J]. 舰船科学技术, 2022, 44(15): 169-172.
ZHANG L J, YIN L Y, JIA Y X. Research on ship target detection method in remote sensing image based on PQFT model[J]. Ship Science and Technology, 2022, 44(15): 169-172.
[19] 赵浩光, 王平, 董超, 等. 结合多尺度视觉显著性的舰船目标检测[J]. 光学精密工程, 2020, 28(6): 1395-1403.
ZHAO H G, WANG P, DONG C, et al. Ship detection based on the multi-scale visual saliency model[J]. Optics and Precision Engineering, 2020, 28(6): 1395-1403.
[20] 何丽, 张红艳, 房婉琳. 融合多尺度边界特征的显著实例分割[J]. 计算机科学与探索, 2022, 16(8): 1865-1876.
HE L, ZHANG H Y, FANG W L. Salient instance segmentation via multiscale boundary characteristic network[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(8): 1865-1876.
[21] TAN M X, LE V Q. EfficientNetV2: smaller models and faster training[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2021.
[22] 余震, 何留杰, 王振飞. 基于中智理论与方向α-均值的图像边缘检测算法[J]. 电子测量与仪器学报, 2020, 32(3): 8-16.
YU Z, HE L J, WANG Z F. Image edge detection based on intelligence theory and direction α-mean[J]. Journal of Electronic Measurement and Instrumentation, 2020, 32(3): 8-16.
[23] QI S, MA J, LIN J, et al. Unsupervised ship detection based on saliency and s-hog descriptor from optical satellite images[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12: 1451-1455.