[学术文献] Regional water resource carrying capacity evaluation based on multi-dimensional precondition cloud and risk matrix coupling model 进入全文
Science of The Total Environment
Water resource carrying capacity modelling is a fundamental task to explore the interaction mechanism between socio-economic development and water resource carrying system. To reasonably quantify regional water resource carrying capacity, firstly, the water resource carrying system was divided into pressure, support and regulation forces subsystems, then the multi-dimensional precondition cloud algorithm was introduced to quantify the belonging degree of single evaluation index, and the comprehensive belonging degree of each sample was further obtained through risk matrix and index weight, and finally the multi-dimensional precondition cloud and risk matrix coupling model (PCRM) was established to recognize carrying grade and reveal carrying mechanism. The application results of PCRM model indicated that water resource carrying capacity in Anhui province, China presented a slightly improving trend in both provincial and city scales during 2005 to 2015. Meanwhile, evaluation result of PCRM model was more approaching to the average characteristic value of different approaches, which indicated that PCRM coupling model is effectively to explore properties of indexes and subsystems of water resource carrying system, and could be further applied in other system evaluation and regulation research fields in the future.
[学术文献] Quantifying smallholder farmers’ managed land use/land cover dynamics and its drivers in contrasting agro-ecological zones of the East African Rift 进入全文
Global Ecology and Conservation
Understanding the relationships between land use land cover (LULC) change and its drivers is essential for designing appropriate strategies for managing and conserving natural resources. This study examined the status, trends, and driving factors of smallholder farmers’ managed land use land cover (LULC) dynamics over the past 30 years (1988e2018) in two contrasting agro-ecological zones (AEZ) of highly-populated Southeastern escarpments of the Ethiopian Rift Valley. Changes in LULC were quantified by integrating field observation, remote-sensing data, and geographic information systems. Landsat images (1988, 2003, and 2018), household surveys (for drivers), and time-series data sources were used to generate the datasets. Results showed that: (1) the total area of 75,246.98 ha (33.79%) was non-linearly changed to various LULC classes. Cultivated land and agroforestry were dominant and increased in humid AEZ from 1988 to 2018. In the same period, waterbody, wetland/marshy, and grass/bushland were dominant and increased in semi-arid AEZ. Cultivated land increased remarkably at the expense of wetland/marshy, woodland, forestland, and grass/bushland in both humid and semi-arid AEZs, and, consequently, bare land was expanded by 1288% in the region. Agroforestry was highly persistent, while forestland was found to be highly susceptible to LULC change. (2) We observed that there was a spatial variation of drivers between humid and semi-arid AEZs. Population growth and landholding size (land fragmentation) were considered to be the main drivers of LULC changes in humid AEZ, while drought/rainfall variability and agricultural investment were the principal drivers in semi-arid AEZ. Ordinary least square (OLS) and binary logistic regression have also confirmed that population density and agricultural investment are the top drivers that significantly influence LULC classes. We conclude that unmanaged population growth, coupled with the continued expansion of cultivated land, have remarkably contributed to the expansion of bare land and decreases vegetation cover. Hence, improving farm-level participatory land-use management is required to recuperate the environment and smallholder farmers’ well-being.
[学术文献] Land-cover classification with high-resolution remote sensing images using transferable deep models 进入全文
Remote Sensing of Environment
In recent years, large amount of high spatial-resolution remote sensing (HRRS) images are available for land-cover mapping. However, due to the complex information brought by the increased spatial resolution and the data disturbances caused by different conditions of image acquisition, it is often difficult to find an efficient method for achieving accurate land-cover classification with high-resolution and heterogeneous remote sensing images. In this paper, we propose a scheme to apply deep model obtained from labeled land-cover dataset to classify unlabeled HRRS images. The main idea is to rely on deep neural networks for presenting the contextual information contained in different types of land-covers and propose a pseudo-labeling and sample selection scheme for improving the transferability of deep models. More precisely, a deep Convolutional Neural Networks (CNNs) is first pre-trained with a well-annotated land-cover dataset, referred to as the source data. Then, given a target image with no labels, the pre-trained CNN model is utilized to classify the image in a patch-wise manner. The patches with high confidence are assigned with pseudo-labels and employed as the queries to retrieve related samples from the source data. The pseudo-labels confirmed with the retrieved results are regarded as supervised information for fine-tuning the pre-trained deep model. To obtain a pixel-wise land-cover classification with the target image, we rely on the fine-tuned CNN and develop a hybrid classification by combining patch-wise classification and hierarchical segmentation. In addition, we create a large-scale land-cover dataset containing 150 Gaofen-2 satellite images for CNN pre-training. Experiments on multi-source HRRS images, including Gaofen-2, Gaofen-1, Jilin-1, Ziyuan-3, Sentinel-2A, and Google Earth platform data, show encouraging results and demonstrate the applicability of the proposed scheme to land-cover classification with multi-source HRRS images.
目前，关于人工智能在农业领域所具备的价值以及将发展到何种程度，尚未达成共识。其预期规模从2017年的2.4亿至5.2亿美元，到2025年的7.90亿至26.28亿美元波动。但可以明确的是，人们期待人工智能在农业上的应用能带来价值上的飞跃。2017年，美国市场约占全球人工智能消费的43%，而欧洲约为23%。 人工智能目前的应用： 关于人工智能在农业领域的应用，众说纷纭。IBM认为人工智能可以通过以下方式服务于农业: 1. 协同物联网技术发挥其最大潜力； 2. 图像识别与洞察； 3. 技能和劳动力； 4. 帮助实现农作物产量的回报最大化； 5. 可对话的农业机器人。