Xiandan Du

In karst mountainous tobacco fields, white plastic mulch, bare rock, and weeds are spatially interspersed and can affect semantic segmentation accuracy in UAV RGB imagery. This study constructed seven background-composition scenarios from UAV RGB orthomosaics acquired in Zhenfeng County, Guizhou Province, China, and evaluated DeepLabV3+, U-Net, U-Net++, and U-Net Former in terms of overall performance, scene-specific accuracy, and cross-scene stability. DeepLabV3+ achieved the highest validation mIoU (0.91) during hyperparameter selection, whereas U-Net performed best on the spatially independent test set, with an mIoU of 0.86, a tobacco-class IoU of 0.74, and an OA of 0.99. Under single-background conditions, bare rock yielded the highest IoU for all four models, while the relative effects of white plastic mulch and weeds were model-dependent. Composite-background effects were non-monotonic: adding weeds to bare-rock scenes reduced IoU by 16.89–28.98%, whereas the three-background scenario did not yield the lowest IoU. U-Net achieved the highest scene-wise IoU in all seven scenarios and the lowest cross-scene coefficient of variation (8.06%). This study provides a reference for sample optimization, model selection, and error diagnosis in complex mountainous tobacco plant identification.

As a core karst region in Southwest China, Guanling County plays a crucial role in regional ecological governance. This study integrates the InVEST model, landscape pattern index analysis, and the MGWR spatial model to systematically explore the dynamic mechanisms of habitat quality in Guanling’s karst mountains. Key findings include: (1) Landscape pattern alterations exhibit significant impacts on habitat quality, characterized by strong spatial heterogeneity; (2) Expansion of forest and grassland effectively buffers the negative effects of construction land expansion, forming an ecological compensation mechanism through enhanced landscape connectivity; (3) Between 2000 and 2020, the proportion of high-importance habitat quality zones increased from 54.79% to 56.16%, with moderate-importance zones stabilizing at approximately 7.80% and general-importance zones growing to 2.46%. The results provide a multi-scale analytical framework for habitat protection and land use optimization in fragile karst ecosystems.

The exploration of the impact of different spatial scales on the low-altitude remote sensing identification of Chinese cabbage (Brassica rapa subsp. Pekinensis) plants offers important theoretical reference value in balancing the accuracy of plant identification with work efficiency. This study focuses on Chinese cabbage plants during the rosette stage; RGB images were obtained by drones at different flight heights (20 m, 30 m, 40 m, 50 m, 60 m, and 70 m). Spectral sampling analysis was conducted on different ground backgrounds to assess their separability. Based on the four commonly used vegetation indices for crop recognition, the Excess Green Index (ExG), Red Green Ratio Index (RGRI), Green Leaf Index (GLI), and Excess Green Minus Excess Red Index (ExG-ExR), the optimal index was selected for extraction. Image processing methods such as frequency domain filtering, threshold segmentation, and morphological filtering were used to reduce the impact of weed and mulch noise on recognition accuracy. The recognition results were vectorized and combined with field data for the statistical verification of accuracy. The research results show that (1) the ExG can effectively distinguish between soil, mulch, and Chinese cabbage plants; (2) images of different spatial resolutions differ in the optimal type of frequency domain filtering and convolution kernel size, and the threshold segmentation effect also varies; (3) as the spatial resolution of the imagery decreases, the optimal window size for morphological filtering also decreases, accordingly; and (4) at a flight height of 30 m to 50 m, the recognition effect is the best, achieving a balance between recognition accuracy and coverage efficiency. The method proposed in this paper is beneficial for agricultural growers and managers in carrying out precision planting management and planting structure optimization analysis and can aid in the timely adjustment of planting density or layout to improve land use efficiency and optimize resource utilization.

Given the complex terrain and economic development status of Guizhou Province, research on tropospheric NO₂ column concentration using satellite remote sensing is still insufficient. Observing the spatial–temporal evolution characteristics of tropospheric NO₂ column concentration can ensure the stable development of air quality. Based on the Google Earth Engine (GEE) platform, NO₂ column concentration data retrieved from Sentinel-5P TROPOMI were analyzed using spatial autocorrelation, hotspot analysis, and geographic detector methods (Geodetector). The results show that NO₂ column concentration in Guizhou Province exhibits seasonal variation, characterized by higher levels in winter and lower levels in summer, with transitional values in spring and autumn. The annual average concentration was highest in 2021 at 3.47 × 10⁻⁵ mol/m² and lowest in 2022 at 2.85 × 10⁻⁵ mol/m². Spatially, NO₂ column concentration displays a distribution pattern of “high in the west, low in the east; high in the north, low in the south”, with significant spatial clustering. The distribution of cold and hot spots aligns with areas of high and low values. NO₂ column concentration is primarily influenced by socio-economic factors, with the interaction between any two factors enhancing the explanatory power of individual factors on NO₂ column concentration.

In order to meet the growing demand for food and achieve food security development goals, contemporary agriculture increasingly depends on plastic coverings such as agricultural plastic films. The remote sensing-based identification of these plastic films has gradually become a necessary tool for agricultural production management and soil pollution prevention. Addressing the challenges posed by the complex terrain and fragmented land parcels in karst mountainous regions, as well as the frequent presence of cloudy and foggy weather conditions, the extraction efficacy of mulching films is compromised. This study utilized a DJI Mavic 2 Pro UAV to capture visible light images in an area with complex terrain features such as peaks and valleys. A plastic film sample dataset was constructed, and the U-Net deep learning model parameters integrated into ArcGIS Pro were continuously modified and optimized to achieve precise plastic film identification. The results are as follows: (1) Sample quantity significantly affects recognition performance. When the sample size is 800, the accuracy of plastic film extraction notably improves, with area accuracy reaching 91%, a patch quantity accuracy of 96.38%, and an IOU and F1-score of 85.89% and 94.20%, respectively, compared to the precision achieved with a sample size of 300; (2) Different learning rates, batch sizes, and iteration numbers have a certain impact on the training effectiveness of the U-Net model. The most suitable model parameters improved the training effectiveness, with the highest training accuracy achieved at a learning rate of 0.001, a batch size of 10, and 25 iterations; (3) Comparative experiments with the Support Vector Machine (SVM) model validate the suitability of U-Net model parameters and sample datasets for precise identification in rugged terrains with fragmented spatial distribution, particularly in karst mountainous regions. This underscores the applicability of the U-Net model in recognizing plastic film coverings in karst mountainous regions, offering valuable insights for agricultural environmental health assessment and green planting management in farmlands.