Deep Convolutional Feature Fusion Model for Multispectral Maritime Imagery Ship Recognition

Combining both visible and infrared object information, multispectral data is a promising source data for automatic maritime ship recognition. In this paper, in order to take advantage of deep convolutional neural network and multispectral data, we model multispectral ship recognition task into a convolutional feature fusion problem, and propose a feature fusion architecture called Hybrid Fusion. We fine-tune the VGG-16 model pre-trained on ImageNet through three channels single spectral image and four channels multispectral images, and use existing regularization techniques to avoid over-fitting problem. Hybrid Fusion as well as the other three feature fusion architectures is investigated. Each fusion architecture consists of visible image and infrared image feature extraction branches, in which the pre-trained and fine-tuned VGG-16 models are taken as feature extractor. In each fusion architecture, image features of two branches are firstly extracted from the same layer or different layers of VGG-16 model. Subsequently, the features extracted from the two branches are flattened and concatenated to produce a multispectral feature vector, which is finally fed into a classifier to achieve ship recognition task. Furthermore, based on these fusion architectures, we also evaluate recognition performance of a feature vector normalization method and three combinations of feature extractors. Experimental results on the visible and infrared ship (VAIS) dataset show that the best Hybrid Fusion achieves 89.6% mean per-class recognition accuracy on daytime paired images and 64.9% on nighttime infrared images, and outperforms the state-of-the-art method by 1.4% and 3.9%, respectively.

A Novel SAR Image Ship Small Targets Detection Method

To satisfy practical requirements of high real-time accuracy and low computational complexity of synthetic aperture radar (SAR) image ship small target detection, this paper proposes a small ship target detection method based on the improved You Only Look Once Version 3 (YOLOv3). The main contributions of this study are threefold. First, the feature extraction network of the original YOLOV3 algorithm is replaced with the VGG16 network convolution layer. Second, general convolution is transformed into depthwise separable convolution, thereby reducing the computational cost of the algorithm. Third, a residual network structure is introduced into the feature extraction network to reuse the shallow target feature information, which enhances the detailed features of the target and ensures the improvement in accuracy of small target detection performance. To evaluate the performance of the proposed method, many experiments are conducted on public SAR image datasets. For ship targets with complex backgrounds and small ship targets in the SAR image, the effectiveness of the proposed algorithm is verified. Results show that the accuracy and recall rate improved by 5.31% and 2.77%, respectively, compared with the original YOLOV3. Furthermore, the proposed model not only significantly reduces the computational effort, but also improves the detection accuracy of ship small target.

Revisiting the dependence of thermocline-dwelling foraminiferal B/Ca on temperature and [CO32-], and its application in reconstruction of the subsurface carbonate system in the tropical western Pacific since 24 ka

The B/Ca ratio of planktonic foraminifer shells has been used as a proxy for reconstructing past ocean carbonate chemistry. However, recent studies have revealed significant uncertainties associated with this proxy, such as whether seawater temperature or [CO^2-3 ] is the dominant control on the partition coefficient (KD) of planktonic foraminiferal B/Ca. To address these uncertainties and thus improve our understanding of the planktonic foraminiferal B/Ca proxy, we analysed B/Ca ratios in the tests of Neogloboquadrina dutertrei (300- 355 μm) and Pulleniatina obliquiloculata (355- 400 μm) in surface sediment samples from the tropical western Pacific and South China Sea. The relationship between these B/Ca ratios and bottom water calcite saturation states (Δ[CO^2-3 ]) is weak, thus suggesting only a small dissolution effect on the B/Ca of the two species. The correlation coefficients (R2) between the B/Ca ratios of N. dutertrei and P. obliquiloculata and environmental parameters (e.g., temperature, salinity, phosphate, DIC and ALK) in the tropical western Pacific and South China Sea are not high enough to justify using B/Ca ratios as a palaeoenvironmental proxy in the study areas. The significant correlation between KD values of N. dutertrei and P. obliquiloculata and carbonate system parameters (e.g.,[CO^2-3 ], DIC, ALK, pH and [HCO^-3 ]) in the study area reflect chemical links between the KD denominator and these variables. Based on our surface sediment calibration, an empirical relationship between the KD of N. dutertrei and temperature is proposed in the tropical western Pacific. We also generated a record of B/Ca ratios in N. dutertrei (300 -355 μm) from Core MD06-3052 in the tropical western Pacific over the past 24 ka to evaluate the application of the revised B/Ca proxy method. Based on the reconstructed empirical relationship for B/Ca and subsurface seawater ALK, we estimated subsurface seawater carbonate system parameters in the tropical western Pacific since 24 ka. In general, the estimated subsurface seawater pH and [CO^2-3 ] show an increase with time, and the record of subsurface seawater pCO2 shows a decrease with time, in the tropical western Pacific over the past 24 ka. The consistent trends in subsurface seawater pCO2 and opal flux during deglaciation may imply that the reported increase in subsurface water pCO2 in the study area was promoted by enhanced upwelling in the Southern Ocean.