Deep-learning-based ship detection methods in Synthetic Aperture Radar (SAR) imagery are a current research hotspot. However, these methods rely on high-quality images as input, and in practical applications, SAR images are interfered with by speckle noise, leading to a decrease in image quality and thus affecting detection accuracy. To address this problem, we propose a unified framework for ship detection that incorporates a despeckling module into the object detection network. This integration is designed to enhance the detection performance, even with low-quality SAR images that are affected by speckle noise. Secondly, we propose a Multi-Scale Window Swin Transformer module. This module is adept at improving image quality by effectively capturing both global and local features of the SAR images. Additionally, recognizing the challenges associated with the scarcity of labeled data in practical scenarios, we employ an unlabeled distillation learning method to train our despeckling module. This technique avoids the need for extensive manual labeling and making efficient use of unlabeled data. We have tested the robustness of our method using public SAR datasets, including SSDD and HRSID, as well as a newly constructed dataset, the RSSDD. The results demonstrate that our method not only achieves a state-of-the-art performance but also excels in conditions with low signal-to-noise ratios.

Understanding spatiotemporal mismatches between urban expansion and population dynamics is essential for guiding sustainable development in rapidly urbanizing regions. Using multi-source nighttime light (NTL) images and global settlement layers, this study investigates the settlement growth pattern and potential spatiotemporal mismatch with population distribution in Guangdong, China, from 1995 to 2019 at a 5-year interval. Specifically, population spatialization in urban and rural areas is separately mapped by adopting a population-based thresholding method, achieving strong agreement with the census record. Our analysis reveals distinct expansion patterns and mismatch conditions across Guangdong’s Core, Belt, and District subzones. The Core and District subzones primarily experienced infilling and edge-expansion urban growth, while the Belt subzone exhibited more dispersed spatial patterns. Notably, only 5 of 21 prefectures exhibited faster population growth than urban expansion, likely due to sustained migration driven by economic opportunities and advanced urbanization. Quantitatively, both urban expansion and population growth followed a Core, Belt, District order. Spatially, population-dominated areas were primarily clustered within 10 km of urban centers, while the District subzone extensively displayed overfilled settlements, indicating low-efficient land use. Temporally, urban growth relative to population in the Core subzone turned from slower pre-2000 to faster post-2000, followed by gradual deceleration, while the Belt subzone maintained balanced growth throughout the study period. The District subzone sustained faster urban growth from 2000 to 2019. Findings of the study provide an important reference for scientific urban planning and sustainable regional development, not only in Guangzhou but other rapidly urbanizing regions globally.