Search

Accurately grasping the distribution and area of cotton for agricultural irrigation scheduling, intensive and efficient management of water resources, and yield estimation in arid and semiarid regions is of great significance. In this paper, taking the Xinjiang Shihezi oasis agriculture region as the study area, extracting the spectroscopic characterization (R, G, B, panchromatic), texture feature (entropy, mean, variance, contrast, homogeneity, angular second moment, correlation, and dissimilarity) and characteristics of vegetation index (normalized difference vegetation index/NDVI, ratio vegetation index/DVI, difference vegetation index/RVI) in the cotton flowering period before and after based on GF-6 image data, four models such as the random forests (RF) and deep learning approach (U-Net, DeepLabV3+ network, Deeplabv3+ model based on attention mechanism) were used to identify cotton and to compare their accuracies. The results show that the deep learning model is better than that of the random forest model. In all the deep learning models with three kinds of feature sets, the recognition accuracy and credibility of the DeepLabV3+ model based on the attention mechanism are the highest, the overall recognition accuracy of cotton is 98.23%, and the kappa coefficient is 96.11. Using the same Deeplabv3+ model based on an attention mechanism with different input feature sets (all features and only spectroscopic characterization), the identification accuracy of the former is much higher than that of the latter. GF-6 satellite image data in the field of crop type recognition has great application potential and prospects.

Climate change has a significant impact on forest ecosystems worldwide, but it is unclear whether forest responses to climate fluctuations are homogeneous across regions. In this study, we investigated the impact of climatic fluctuations on forest growth in a complex terrain, in Anhui Province of China, using Enhanced Vegetation Index (EVI) data from the Moderate-Resolution Imaging Spectroradiometer (MODIS), while considering the impact of terrain characteristics and forest types. Our regional-scale analysis found that the forest response to climatic drivers in Anhui Province is not homogeneous, with only 69% of the forest area driven by temperature (TEM), while 11% is precipitation (PRE) driven and 20% is solar radiation (SWD) driven. We also found with random forest models that terrain traits (elevation and slope) contributed significantly (29.47% and 27.96%) to the spatial heterogeneity of forest response to climatic drivers, with higher elevation associated with a stronger positive correlation between the EVI and temperature (p < 0.001), a weaker positive correlation between the EVI with precipitation (p < 0.001), and a stronger negative correlation between the EVI with solar radiation (p < 0.01), while forest type contributed the least (4.21%). Our results also imply that in a warmer and dryer climate, some forest patches may switch from TEM driven to PRE driven, which could lead to a decrease in forest productivity, instead of an increase as predicted by existing climate models. These results highlight the heterogeneous response of forests within close vicinity to climate fluctuations in a complex terrain, which has important implications for climate-related risk assessments and local forest management.

Cloud and snow identification in remote sensing images is critical for snow mapping and snow hydrology research. Aimed at the problem that the semantic segmentation model is prone to producing blurred boundaries, slicing traces and isolated small patches for cloud and snow identification in high-resolution remote sensing images, the feasibility of combining DeepLab v3+ and conditional random field (CRF) models for cloud and snow identification based on GF-1 WFV images is studied. For GF-1 WFV images, the model training and testing experiments under the conditions of different sample numbers, sample sizes and loss functions are compared. The results show that, firstly, when the number of samples is 10,000, the sample size is 256 × 256, and the loss function is the Focal function, the model accuracy is the optimal and the Mean Intersection over Union (MIoU) and the Mean Pixel Accuracy (MPA) reach 0.816 and 0.918, respectively. Secondly, after post-processing with the CRF model, the MIoU and the MPA are improved to 0.836 and 0.941, respectively, compared with those without post-processing. Moreover, the misclassifications such as blurred boundaries, slicing traces and isolated small patches are significantly reduced, which indicates that the combination of the DeepLab v3+ and CRF models has high accuracy and strong feasibility for cloud and snow identification in high-resolution remote sensing images. The conclusions can provide a reference for high-resolution snow mapping and hydrology applications using deep learning models.

The chlorophyll content of leaves is an important indicator of plant environmental stress, photosynthetic capacity, and is widely used to diagnose the growth and health status of vegetation. Traditional chlorophyll content inversion is based on the vegetation index under pure species, which rarely considers the impact of interspecific competition and species mixture on the inversion accuracy. To solve these limitations, the harmonic analysis (HA) and the Hilbert–Huang transform (HHT) were introduced to obtain the frequency index, which were combined with spectral index as the input parameters to estimate chlorophyll content based on the unmanned aerial vehicle (UAV) image. The research results indicated that: (1) Based on a comparison of the model accuracy for three different types of indices in the same period, the estimation accuracy of the pure spectral index was the lowest, followed by that of the frequency index, whereas the mixed index estimation effect was the best. (2) The estimation accuracy in November was lower than that in other months; the pure spectral index coefficient of determination (R2) was only 0.5208, and the root–mean–square error (RMSE) was 4.2144. The estimation effect in September was the best. The model R2 under the mixed index reached 0.8283, and the RMSE was 2.0907. (3) The canopy chlorophyll content (CCC) estimation under the frequency domain index was generally better than that of the pure spectral index, indicating that the frequency information was more sensitive to subtle differences in the spectrum of mixed vegetation. These research results show that the combination of spectral and frequency information can effectively improve the mapping accuracy of the chlorophyll content, and provid a theoretical basis and technology for monitoring the chlorophyll content of mixed vegetation in wetlands.