شبکه عصبی عمیق مبتنی بر کالیبراسیون مجدد
Abstract
I. Introduction
II. Methods
III. Experiments
IV. Discussion
V. Conclusion
Authors
Figures
References
Abstract
Ship detection plays an important role in synthetic aperture radar (SAR) image interpretation. However, there are still some difficulties in SAR ship detection. First, ships often have a large aspect ratio and arbitrary directionality in SAR images. Traditional detection algorithms can cause the detection area to be redundant, which makes it difficult to accurately locate the target in complex scenes. Second, ships in ports are often densely arranged, and the effective identification of densely arranged ships is complicated. Finally, ships in SAR images exist at a variety of scales due to the multiresolution imaging modes used and ship shape variations, which pose a considerable challenge for ship detection. To solve the above problems, we propose a multiscale adaptive recalibration network (MSARN) to detect multiscale and arbitrarily oriented ships in complex scenarios. The recalibration of the extracted multiscale features through global information increases the sensitivity of the network to the target angle, thereby increasing the accuracy of positioning. In particular, we designed a pyramid anchor and a loss function to match the rotated target. In addition, we modified the rotation non-maximum suppression (RNMS) method to solve the problem of the large overlap ratio of the detection box. The proposed model combines the positioning advantage of rotation detection with the speed advantage of a single-stage framework. Experiments show that based on the SAR rotation ship detection (SRSD) data set, the proposed algorithm has a faster detection speed and higher accuracy than some state-of-the-art methods.
Introduction
Synthetic aperture radar (SAR) has been widely used in civil remote sensing surveying and military reconnaissance due to its independence from solar illumination and ability to provide images in all-weather operating conditions [1]–[4]. In recent years, SARs such as TerraSAR-X, RADARSAT-2, and Sentinel-1 have rapidly developed, which has greatly promoted research on SAR image ship detection methods [5]–[7]. The constant false alarm rate (CFAR) and its various derivative algorithms are widely used in SAR image ship detection [8]–[10]. This type of algorithm is based on a statistical model of the contrast information, which can automatically adjust thresholds to suit different ocean backgrounds while maintaining the required performance. However, the algorithm modelling process is complicated, and ship detection in complex backgrounds cannot achieve the desired effect. A deep convolutional neural network (DCNN) can automatically learn the structural features of a target and have been rapidly developed in the field of computer vision. A series of object detection algorithms based on DCNN was proposed. The detection algorithms based on DCNN mainly include two types: two-stage detection algorithms, such as the Faster R-CNN [11], and single-stage detection algorithms, such as SSD [12], YOLO [13]–[15], and RFBNet [16].