Side-scan sonar is widely used in underwater rescue and the detection of undersea targets, such as shipwrecks, aircraft crashes, etc. Automatic object classification plays an important role in the rescue process to reduce the workload of staff and subjective errors caused by visual fatigue. However, the application of automatic object classification in side-scan sonar images is still lacking, which is due to a lack of datasets and the small number of image samples containing specific target objects. Secondly, the real data of side-scan sonar images are unbalanced. Therefore, a side-scan sonar image classification method based on synthetic data and transfer learning is proposed in this paper. In this method, optical images are used as inputs and the style transfer network is employed to simulate the side-scan sonar image to generate “simulated side-scan sonar images”; meanwhile, a convolutional neural network pre-trained on ImageNet is introduced for classification. In this paper, we experimentally demonstrate that the maximum accuracy of target classification is up to 97.32% by fine-tuning the pre-trained convolutional neural network using a training set incorporating “simulated side-scan sonar images”. The results show that the classification accuracy can be effectively improved by combining a pre-trained convolutional neural network and “similar side-scan sonar images”.
The Yellow River Basin (YRB) is a critical ecological zone in China now confronting growing tensions between land conservation and development. This study combines land use, climate, and socio-economic data with spatial–statistical models (GeoDetector [GD]–Patch-generating Land Use Simulation [PLUS]–Integrated Valuation of Ecosystem Services and Trade-Offs [InVEST]) to analyze land use changes (2000–2020), evaluate habitat quality, and simulate scenarios to 2040. Key results include the following: (1) Farmland was decreased by the conversion to forests (+3475 km2) and grasslands (+4522 km2), while construction land expanded rapidly (+11,166 km2); (2) the population and Gross Domestic Product (GDP) pressures drove the farmland loss (q = 0.148 for population, q = 0.129 for GDP), while synergies between evapotranspiration (ET) and the Normalized Difference Vegetation Index (NDVI) promoted forest/grassland recovery (q = 0.155); and (3) ecological protection scenarios increased the grassland area by 12.94% but restricted the construction land growth (−13.84%), with persistent unused land (>3.61% in Inner Mongolia) indicating arid-zone risks. The Habitat Quality-Autocorrelated Coupling Index (HQACI) declined from 0.373 (2020) to 0.345–0.349 (2040), which was linked to drought, groundwater loss, and urban expansion. Proposed strategies including riparian corridor protection, adaptive urban zoning, and gradient-based restoration aim to balance ecological and developmental needs, supporting spatial planning and enhancing the basin-wide habitat quality.
As one of the most important greenhouse gases, carbon dioxide (CO2) exhibits spatiotemporal variations that directly affect the accuracy of global carbon inventories. In recent years, multiple satellites have successively been deployed for observing the column-averaged CO2 dry-air mole fraction (XCO2). However, these satellites perform quite differently, so it is crucial to evaluate their XCO2 products systematically for both scientific and practical reasons. Most existing studies rely on ground-based observations or the CarbonTracker (CT) model data as reference benchmarks. Nevertheless, because ground-based stations are sparsely distributed and model data are subject to prior errors, biases may be introduced into the evaluation results. In contrast, the Three-Cornered Hat (TCH) method can estimate the relative errors of multi-source data without true values. Based on this, the current study systematically evaluates the XCO2 products of the four following satellites—Greenhouse Gases Observing Satellite (GOSAT), GOSAT-2, Orbiting Carbon Observatory 2 (OCO-2), and OCO-3—over China by integrating the TCH method, ground-based observations and CarbonTracker model data. The results show that the monthly coverage of the four satellite XCO2 products in China is limited. In terms of overall performance, the OCO-series outperforms the GOSAT-series, with OCO-3 showing the relatively best performance. Additionally, the TCH method proves to be applicable and reliable for uncertainty analysis of XCO2 data. This study provides a new perspective for the quality grading and fusion application of multi-source satellite XCO2 data, and is of great significance for carbon assimilation models.
Understanding the spatiotemporal dynamics of land-use and cover change (LUCC) and ecosystem service (ES) responses is essential for assessing ecological functions in regional landscapes. However, conventional LUCC simulations often rely on historical trends and inadequately represent the spatially heterogeneous effects of top-down policy constraints. Taking the Henan section of the Yellow River Basin (HYRB) as a case study, we developed a policy-to-rule framework that translated ecological redlines, urban development boundaries, and restoration requirements into explicit spatial constraints and land-use transition rules in the PLUS model. A policy-constrained High-Quality Development Scenario (HQDS) was established, with the Natural Growth Scenario (NGS) as a reference. Five ESs were assessed using InVEST from 1985 to 2050, and the results were integrated with the Minimum Cumulative Resistance (MCR) model and circuit theory to construct an ecological security pattern (ESP). Historical reconstruction of the 2022 land-use pattern achieved an overall accuracy of 90.18% and a Kappa coefficient of 86.39%. The five ESs remained relatively stable overall: water yield, soil conservation, and the sediment-related indicator increased, whereas habitat quality and carbon storage declined slightly. Ecological source areas expanded from 7140.54 km2 in 1985 to 12,039.17 km2 under the HQDS in 2050, a 68.6% increase. Compared with the NGS, the HQDS increased source areas by 562.42 km2 (4.9%), reduced ecological corridors from 26 to 24, and increased their total length from 1068.89 to 1099.61 km. These differences represent the projected, scenario-conditioned consequences of the specified policy constraints and provide quantitative decision support for future ecological management.