Monitoring the quality of the urban ecological environment has become one of the important elements of promoting a sustainable urban development. The remote sensing ecological index (RSEI) provides a new research direction in urban ecological environment monitoring, combined with remote sensing. However, by using the principal component analysis method in RSEI, the calculation results are complicated and the workload is huge. To effectively assess the urban ecological environment, an improved remote sensing ecological index (IRSEI) was created to improve the ease of data use by using the entropy weighting method with spatiotemporal characteristics and seasonal variations. Furthermore, a geographically weighted regression model was used to quantify the impact of human activities on the urban ecological environment quality. The results showed that the IRSEI could provide a new method for monitoring the urban ecological environment quality, which makes the work easier while ensuring the validity of the results. It was concluded that (1) seasonal differences in the ecological quality of the study area were evident in the IRSEI model and the overall ecological environment quality of Jining City had been on an upward trend in the past 20 years; (2) the ecological quality in the study area was unevenly distributed spatially, with the southwestern part being better than the northeastern part, and the ecological grade being predominantly between moderate and good; and (3) the spatial aggregation effect of the IRSEI was increasing with time. The geographically weighted regression (GWR) revealed the influence of human activities on the ecological environment quality, among which economic level was positively related to ecological improvement, but the population density and night light index were negatively related to improvements in the ecological environment; road network density only showed a negative correlation in 2020. As Jining urbanizes, attention should be paid to protecting the built environment and population distribution.
Deep learning algorithms offer an effective solution to the inefficiencies and poor results of traditional methods for building a footprint extraction from high-resolution remote sensing imagery. However, the heterogeneous shapes and sizes of buildings render local extraction vulnerable to the influence of intricate backgrounds or scenes, culminating in intra-class inconsistency and inaccurate segmentation outcomes. Moreover, the methods for extracting buildings from very high-resolution (VHR) images at present often lose spatial texture information during down-sampling, leading to problems, such as blurry image boundaries or object sticking. To solve these problems, we propose the multi-scale boundary-refined HRNet (MBR-HRNet) model, which preserves detailed boundary features for accurate building segmentation. The boundary refinement module (BRM) enhances the accuracy of small buildings and boundary extraction in the building segmentation network by integrating edge information learning into a separate branch. Additionally, the multi-scale context fusion module integrates feature information of different scales, enhancing the accuracy of the final predicted image. Experiments on WHU and Massachusetts building datasets have shown that MBR-HRNet outperforms other advanced semantic segmentation models, achieving the highest intersection over union results of 91.31% and 70.97%, respectively.
Research on the spatiotemporal changes in land use/cover (LUC) and carbon storage (CS) in the region of the Taihang Mountains in various developmental scenarios can provide significant guidance for optimizing the structure of LUC and formulating ecologically friendly economic development policies. We employed the PLUS and InVEST models to study change in LUC and CS in the Taihang Mountains from 1990 to 2020. Based on these results, we established three distinct development scenarios: a business-as-usual development scenario, a cropland protection scenario, and an ecological conservation scenario. Based on these three developmental scenarios, we simulated the spatiotemporal changes in LUC and CS in the Taihang Mountains in 2035. The results indicate that: (1) from 1990 to 2020, the CS in the Taihang Mountains increased from 1575.91 Tg to 1598.57 Tg, with a growth rate of approximately 1.44%. The primary source of this growth is attributed to the expansion of forests. (2) In the business-as-usual development scenario, the growth rate of CS in the Taihang Mountains was approximately 0.45%, indicating a slowdown in the trend. This suggests that economic development has the consequences of aggravating human–land conflicts, leading to a deceleration in the growth of CS. (3) In the cropland protection scenario, the increase in the CS in the Taihang Mountains was similar to the CS increase in the business-as-usual development scenario. However, the expansion of cropland dominated by impermeable surfaces, which indicates economic development, was considerably constrained in this scenario. (4) In the ecological conservation scenario, the increase in carbon storage in the Taihang Mountains was 1.16%, which is the fastest among all three scenarios. At the same time, there was a certain degree of development of impermeable surfaces, achieving a balance between economic development and ecological conservation.
1 October 2022 in Journal of Applied Remote Sensing
Convolutional neural networks (CNNs) are very important deep neural networks for analyzing visual imagery. However, most CNN-based methods have the problem of over-smoothing at boundaries, which is unfavorable for hyperspectral image classification. To address this problem, a spectral-spatial multiscale residual network (SSMRN) by fusing two separate deep spectral features and deep spatial features is proposed to significantly reduce over-smoothing and effectively learn the features of objects. In the implementation of the SSMRN, a multiscale residual convolutional neural network is proposed as a spatial feature extractor and a band grouping-based bi-directional gated recurrent unit is utilized as a spectral feature extractor. Considering that the importance of spectral and spatial features may vary depending on the spatial resolution of images, we combine both features with two weighting factors with different initial values that can be adaptively adjusted during the network training. To evaluate the effectiveness of the SSMRN, extensive experiments are conducted on public benchmark data sets. The proposed method can retain the detailed boundary of different objects and yield competitive results compared with several state-of-the-art methods.
In recent years, rapid urbanization in the Southern Jiangsu urban area has significantly altered land use patterns, impacting local ecosystem service value (ESV). This study examines land use changes (LUCs) from 2000 to 2020 and assesses ESV changes with the equivalent factor method. The PLUS model is also used to project future LUCs in line with the “Land and Space Planning of Jiangsu Province (2021–2035)”. The simulation considers scenarios for natural development (NDS), ecological protection (EPS), farmland protection (FPS), and comprehensive development (CDS) for the year 2035. The research findings show that over the past two decades, urbanized areas have persistently grown, whereas there has been a steady decline in the extents of farmland, forested areas, and grasslands. The total ESV has shown a decreasing trend, with the spatial distribution of ESV exhibiting higher values in the eastern regions and lower values in the west. The relationship between land use and ESV is negative, as evidenced by both the comprehensive land use (LA) and the Moran’s I index. Scenario predictions reveal that farmland, forest land, and grassland show a decreasing trend under the NDS, while it is the opposite under the EPS. The FPS is predominant in farmland area, and all land types in the CDS comply with national land planning requirements. Overall, the research findings reveal and predict the influence of LUCs on ESV in the study area under four scenarios, which is beneficial optimizing regional resource planning schemes and formulating ecological compensation policies.
Reliable information of land cover dynamics in dryland cities is crucial for understanding the anthropogenic impacts on fragile environments. However, reduced classification accuracy of dryland cities often occurs in global land cover data. Although many advanced classification techniques (i.e., convolutional neural networks (CNN)) have been intensively applied to classify urban land cover because of their excellent performance, specific classification models focusing on typical dryland cities are still scarce. This is mainly attributed to the similar features between urban and non-urban areas, as well as the insufficient training samples in this specific region. To fill this gap, this study trained a CNN model to improve the urban land classification accuracy for seven dryland cities based on rigorous training sample selection. The assessment showed that our proposed model performed with higher overall accuracy (92.63%) than several emerging land cover products, including Esri 2020 Land Cover (75.55%), GlobeLand30 (73.24%), GLC_FCS30-2020 (69.68%), ESA WorldCover2020 (64.38%), and FROM-GLC 2017v1 (61.13%). In addition, the classification accuracy of the dominant land types in the CNN-classified data exceeded the selected products. This encouraging finding demonstrates that our proposed architecture is a promising solution for improving dryland urban land classification accuracy and compensating the deficiency of large-scale land cover mapping.