With the development of deep learning techniques in the field of remote sensing change detection, many change detection algorithms based on convolutional neural networks (CNNs) and nonlocal self-attention (NLSA) mechanisms have been widely used and have obtained good detection accuracy. However, these methods mainly extract semantic features on images from different periods without taking into account the temporal dependence between these features. This will lead to more “pseudo-change” in complex scenes. In this paper, we propose a network architecture named UVACD for bitemporal image change detection. The network combines a CNNs extraction backbone for extracting high-level semantic information with a visual transformer. Here, visual transformer constructs change intensity tokens to complete the temporal information interaction and suppress irrelevant information weights to help extract more distinguishable change features. Our network is validated and tested on both the LEVIR-CD and WHU datasets. For the LEVIR-CD dataset, we achieve an intersection over union (IoU) of 0.8398 and an F1 score of 0.9130. For the WHU dataset, we achieve an IoU of 0.8664 and an F1 score of 0.9284. The experimental results show that the proposed method outperforms some previous state of the art change detection methods.

Atmospheric particulate matter (PM) is a major air pollutant. PM_2.5 and PM₁₀ pose particularly serious threats to the ecological environment and human health. Vegetation plays an important role in reducing the concentration of particles. Based on a long time series of air quality, meteorological, and vegetation coverage data in the Beijing–Tianjin–Hebei (BTH) region, the present paper evaluated the influence at the overall and built-up area scales and quantified the process involved in the dry settlement of particles on vegetation based on a mathematical model. The experimental results showed that (1) the total amounts of PM₁₀ reduced by vegetation in the BTH area were 505,200 t, 465,500 t, 477,200 t and 396,500 t in 2015, 2016, 2017 and 2018, respectively, and the total amount of PM_2.5 was reduced by 19,400 t, 19,200 t, 16,400 t and 12,700 t, respectively. The annual reduction in PM₁₀ and PM_2.5 from 2015 to 2018 by vegetation in the BTH region showed a downwards trend, and the annual reduction was mainly caused by the significant decrease in PM concentration. (2) More than 80% of the reduction in annual yield was concentrated in May–September, and a large leaf area was the main reason for the largest yield reduction in the growing season. The efficiency of PM reduction in forestland was approximately five–seven times that in grassland, and the deciduous broad-leaved forest was the main driver of this reduction in each forest. (3) The reduction in PM₁₀ by vegetation was approximately 30 times that of PM_2.5. However, the reduction in PM_2.5 by vegetation should not be ignored because PM_2.5 has a stronger correlation with human production and living activities. Increasing the area and density of green space via afforestation, returning farmland to forest and giving full play to the self-purification function of green spaces are very important to reducing and controlling the concentration of PM.

Based on measurement data from air quality monitoring stations, the spatio-temporal characteristics of the concentrations of particles with aerodynamic equivalent diameters smaller than 2.5 and 10 μm (PM_2.5 and PM₁₀, respectively) in the Beijing–Tianjin–Hebei (BTH) region from 2015 to 2018 were analysed at yearly, seasonal, monthly, daily and hourly scales. The results indicated that (1) from 2015 to 2018, the annual average values of PM_2.5 and PM₁₀ concentrations and the PM_2.5/PM₁₀ ratio in the study area decreased each year; (2) the particulate matter (PM) concentration in winter was significantly higher than that in summer, and the PM_2.5/PM₁₀ ratio was highest in winter and lowest in spring; (3) the PM_2.5 and PM₁₀ concentrations exhibited a pattern of double peaks and valleys throughout the day, reaching peak values at night and in the morning and valleys in the morning and afternoon; and (4) with the use of an improved sine function to simulate the change trend of the monthly mean PM concentration, the fitting R² values for PM_2.5 and PM₁₀ in the whole study area were 0.74 and 0.58, respectively. Moreover, the high-value duration was shorter, the low-value duration was longer, and the concentration decrease rate was slower than the increase rate.

The increasing frequency of human activities has accelerated changes in land use types and consequently affected the atmospheric environment. In this manuscript, we analyze the relationships between the particulate matter concentration and land use changes in the Beijing–Tianjin–Hebei (BTH) region, China, from 2015 to 2018. The experimental results indicate that (1) an improved sine function model can suitably fit the periodic changes in the particulate matter concentration, with the average R² value increasing to 0.65 from the traditional model value of 0.49, while each model coefficient effectively estimates the change characteristics of each stage. (2) Among all land use types, the particulate matter concentrations in construction land and farmland are high, with a large annual difference between high and low values. The concentration decreases slowly in spring and summer but increases rapidly in autumn and winter. The concentrations in forestland and grassland are the lowest; the difference between high and low values is small for these land use types, and the concentration fluctuation pattern is relatively uniform. Natural sources greatly influence the concentration fluctuations, among which frequent dusty weather conditions in spring impose a greater influence on forestland and grassland than on the other land use types. (3) The landscape pattern of land use exerts a significant influence on the particulate matter concentration. Generally, the lower the aggregation degree of patches is, the higher the fragmentation degree is, the more complex the shape is, the higher the landscape abundance is, and the lower the particulate matter concentration is. The higher the construction land concentration is, the more easily emission sources can be aggregated to increase the particulate matter concentration. However, when forestland areas are suitably connected, this land use type can play a notable role in inhibiting particulate matter concentration aggravation. This conclusion is of great relevance to urban land use planning and sustainable development.

As important green energy, wind power and photovoltaic power have great development prospects. The suitability evaluation of wind and solar power plants is a popular research field, which is related to the sustainable and healthy development of wind and solar power generation. In this paper, based on multiple dimensions such as land types, climatic conditions, topographic features and policy environment, we selected 10 indicators and combined (analytical hierarchy process) AHP method to build a suitability assessment model for evaluating the suitability of solar and wind power in Inner Mongolia, China. The findings revealed that, Inner Mongolia has a great potential to generate wind and solar electricity, for wind power, the category of ‘excellent’ regions covers 83855 km² and represents 7.10% of the total surface area; for solar power, 7.66% (nearly 90420 km²) are classified as ‘excellent’. The suitability of both solar and wind energy in the western region is considered to have the most suitable development region, parts of Alxa League and Bayannur City have great potential for combined wind and solar power generation. The research results could provide important technical and data support for capacity evaluation and power station location decision.

Remote sensing change detection (CD) identifies changes in each pixel of certain classes of interest from a set of aligned image pairs. It is challenging to accurately identify natural changes in feature categories due to unstructured and temporal changes. This research proposed an effective bi-temporal remote sensing CD comprising an encoder that could extract multiscale features, a decoder that focused on semantic alignment between temporal features, and a classification head. In the decoder, we constructed a new convolutional attention structure based on pre-generation of depthwise-separable change-salient maps (PDACN) that could reduce the attention of the network on unchanged regions and thus reduce the potential pseudo-variation in the data sources caused by semantic differences in illumination and subtle alignment differences. To demonstrate the effectiveness of the PDA attention structure, we designed a lightweight network structure for encoders under both convolution-based and transformer architectures. The experiments were conducted on a single-building CD dataset (LEVIR-CD) and a more complex multivariate change type dataset (SYSU-CD). The results showed that our PDA attention structure generated more discriminative change variance information while the entire network model obtained the best performance results with the same level of network model parameters in the transformer architecture. For LEVIR-CD, we achieved an intersection over union (IoU) of 0.8492 and an F1 score of 0.9185. For SYSU-CD, we obtained an IoU of 0.7028 and an F1 score of 0.8255. The experimental results showed that the method proposed in this paper was superior to some current state-of-the-art CD methods.