scholarly journals Improving Parcel-Level Mapping of Smallholder Crops from VHSR Imagery: An Ensemble Machine-Learning-Based Framework

2021 ◽  
Vol 13 (11) ◽  
pp. 2146
Author(s):  
Peng Zhang ◽  
Shougeng Hu ◽  
Weidong Li ◽  
Chuanrong Zhang ◽  
Peikun Cheng
Keyword(s):  
Machine Learning ◽  
Land Use ◽  
Precision Agriculture ◽  
Agricultural Land ◽  
Spatial Information ◽  
Classification Performance ◽  
Central China ◽  
Ensemble Machine Learning ◽  
Land Use Types ◽  
Crop Mapping

Explicit spatial information about crop types on smallholder farms is important for the development of local precision agriculture. However, due to highly fragmented and heterogeneous cropland landscapes, fine-scale mapping of smallholder crops, based on low- and medium-resolution satellite images and relying on a single machine learning (ML) classifier, generally fails to achieve satisfactory performance. This paper develops an ensemble ML-based framework to improve the accuracy of parcel-level smallholder crop mapping from very high spatial resolution (VHSR) images. A typical smallholder agricultural area in central China covered by WorldView-2 images is selected to demonstrate our approach. This approach involves the task of distinguishing eight crop-level agricultural land use types. To this end, six widely used individual ML classifiers are evaluated. We further improved their performance by independently implementing bagging and stacking ensemble learning (EL) techniques. The results show that the bagging models improved the performance of unstable classifiers, but these improvements are limited. In contrast, the stacking models perform better, and the Stacking #2 model (overall accuracy = 83.91%, kappa = 0.812), which integrates the three best-performing individual classifiers, performs the best of all of the built models and improves the classwise accuracy of almost all of the land use types. Since classification performance can be significantly improved without adding costly data collection, stacking-ensemble mapping approaches are valuable for the spatial management of complex agricultural areas. We also demonstrate that using geometric and textural features extracted from VHSR images can improve the accuracy of parcel-level smallholder crop mapping. The proposed framework shows the great potential of combining EL technology with VHSR imagery for accurate mapping of smallholder crops, which could facilitate the development of parcel-level crop identification systems in countries dominated by smallholder agriculture.

scholarly journals Responses of the electron transfer capacity of soil humic substances to agricultural land-use types

RSC Advances ◽  
2018 ◽  
Vol 8 (57) ◽  
pp. 32588-32596 ◽  
Author(s):  
Beidou Xi ◽  
Zhurui Tang ◽  
Jie Jiang ◽  
Wenbing Tan ◽  
Caihong Huang ◽  
...  
Keyword(s):  
Land Use ◽  
Electron Transfer ◽  
Humic Substances ◽  
Agricultural Land ◽  
Agricultural Land Use ◽  
Chemical Structures ◽  
Land Use Types ◽  
Electron Transfer Capacity

Agricultural land-use types could affect the transformation and decomposition of HS in soils, and thus further change the intrinsic chemical structures associated with ETC.

scholarly journals Dynamics of forage and land cover changes in Teltele district of Borana rangelands, southern Ethiopia: using geospatial and field survey data

BMC Ecology ◽  
2020 ◽  
Vol 20 (1) ◽  
Author(s):  
Yeneayehu Fenetahun ◽  
Wang Yong-dong ◽  
Yuan You ◽  
Xu Xinwen
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Biomass Production ◽  
Agricultural Land ◽  
Southern Ethiopia ◽  
Forest Land ◽  
Lulc Change ◽  
Grass Land ◽  
Land Use Types ◽  
Bare Land

Abstract Background The gradual conversion of rangelands into other land use types is one of the main challenges affecting the sustainable management of rangelands in Teltele. This study aimed to examine the changes, drivers, trends in land use and land cover (LULC), to determine the link between the Normalized Difference Vegetation Index (NDVI) and forage biomass and the associated impacts of forage biomass production dynamics on the Teltele rangelands in Southern Ethiopia. A Combination of remote sensing data, field interviews, discussion and observations data were used to examine the dynamics of LULC between 1992 and 2019 and forage biomass production. Results The result indicate that there is a marked increase in farm land (35.3%), bare land (13.8%) and shrub land (4.8%), while the reduction found in grass land (54.5%), wet land (69.3%) and forest land (10.5%). The larger change in land observed in both grassland and wetland part was observed during the period from 1995–2000 and 2015–2019, this is due to climate change impact (El-Niño) happened in Teltele rangeland during the year 1999 and 2016 respectively. The quantity of forage in different land use/cover types, grass land had the highest average amount of forage biomass of 2092.3 kg/ha, followed by wetland with 1231 kg/ha, forest land with 1191.3 kg/ha, shrub land with 180 kg/ha, agricultural land with 139.5 kg/ha and bare land with 58.1 kg/ha. Conclusions The significant linkage observed between NDVI and LULC change types (when a high NDVI value, the LULC changes also shows positive value or an increasing trend). In addition, NDVI value directly related to the greenness status of vegetation occurred on each LULC change types and its value directly linkage forage biomass production pattern with grassland land use types. 64.8% (grass land), 43.3% (agricultural land), 75.1% (forest land), 50.6% (shrub land), 80.5% (bare land) and 75.5% (wet land) more or higher dry biomass production in the wet season compared to the dry season.

scholarly journals Spatial Patterns in Baseflow Mean Response Time across a Watershed in the Loess Plateau: Linkage with Land-Use Types

2020 ◽  
Vol 66 (3) ◽  
pp. 382-391 ◽  
Author(s):  
Xu-dong Huang ◽  
Dong Wang ◽  
Pei-pei Han ◽  
Wen-chuan Wang ◽  
Qing-jie Li ◽  
...  
Keyword(s):  
Land Use ◽  
Response Time ◽  
Overland Flow ◽  
Agricultural Land ◽  
Least Squares Regression ◽  
Unit Hydrograph ◽  
Mean Response Time ◽  
Impervious Area ◽  
Land Use Types ◽  
Instantaneous Unit Hydrograph

Abstract Understanding the relation between land-use types and baseflow mean response time (BMRT) is important to explore the response mechanism of baseflow processes in watersheds. BMRT was determined using an instantaneous unit hydrograph. The instantaneous unit hydrograph parameters were estimated by autocorrelation functions. The relative importance of land-use types in determining BMRT dynamics was assessed by hydrological model and partial least-squares regression. Our study suggests greater effects of urban area on BMRT than the effects of forest and agricultural land. This may be because the urban interception impervious area may impede baseflow generation over a short timescale. The effects of agricultural land are greater than those of forest in areas with steeper hillslopes, but lower than those of the forest in areas with more plains, reflecting the varied ability of forest and agricultural lands with different topography to hinder overland flow. Variations of BMRT are strongly linked to land use in the watershed. Overall, our study provides insight into the BMRT and dominant factors of land-use types in watersheds, planning of sustainable water resource use, and ecological protection in watersheds.

Evaluation of soil quality in different land uses in the Mengzi Gabin Basin, Southwest of China

2020 ◽  
Author(s):  
Liqun Tang ◽  
Zhijie Shan ◽  
Yang Yu
Keyword(s):  
Land Use ◽  
Organic Matter ◽  
Soil Organic Matter ◽  
Soil Quality ◽  
Total Nitrogen ◽  
Total Phosphorus ◽  
Agricultural Land ◽  
Organic Matter Content ◽  
Land Use Types ◽  
Total Potassium

<p>Re-vegetation has been widely carried out to prevent land degradation, reduce soil erosion, and improve soil quality. In order to investigate the characteristics of soil nutrients content in different land use types of karst gabin basin, soil organic matter, soil total nitrogen, soil total phosphorus, soil total potassium, soil pH, and soil texture in woodland, agricultural land, orchard, and grassland were surveyed in Mengzi Gabin Basin, Southwest of China. The difference of soil indicators between vegetation types was analyzed, and soil fertility quality of four land use types was comprehensively evaluated by the soil quality index (SQI). The results showed that land use significantly affected soil organic matter content. Soil organic matter content was the highest in grassland, followed by agricultural land and forest land, while orchard was lowest. There was a significant difference in soil total nitrogen content between different land uses. The total nitrogen content in farmland soil was the highest, followed by grassland and woodland, and the lowest in the orchard. Woodand had the highest total potassium content and the lowest total phosphorus content. The grassland soil had the highest total phosphorus content and the lowest total potassium content. pH value in the four land use types was acidic, ranged from 5.82 to 6.67. The soil quality index showed that woodland had the highest soil fertility quality. The results of the study could provide the basis of soil nutrients variation and status in Gabin basin, and also provides support for evaluating the soil improvements during vegetation restoration in fragile Karst ecosystems.</p>

Validation of SWEEP for Contrasting Agricultural Land Use Types in the Tarim Basin

Soil Science ◽  
2014 ◽  
Vol 179 (9) ◽  
pp. 433-445 ◽  
Author(s):  
Huawei Pi ◽  
Gary Feng ◽  
Brenton S. Sharratt ◽  
Xinhu Li ◽  
Zehao Zheng
Keyword(s):  
Land Use ◽  
Tarim Basin ◽  
Agricultural Land ◽  
Agricultural Land Use ◽  
Land Use Types

Agricultural land use change impacts soil CO 2 emission and its 13 C-isotopic signature in central China

2018 ◽  
Vol 177 ◽  
pp. 105-112 ◽  
Author(s):  
Qian Zhang ◽  
Junjun Wu ◽  
Yao Lei ◽  
Fan Yang ◽  
Dandan Zhang ◽  
...  
Keyword(s):  
Land Use ◽  
Land Use Change ◽  
Agricultural Land ◽  
Isotopic Signature ◽  
Central China ◽  
Agricultural Land Use

scholarly journals Relationship Study on Land Use Spatial Distribution Structure and Energy-Related Carbon Emission Intensity in Different Land Use Types of Guangdong, China, 1996–2008

2013 ◽  
Vol 2013 ◽  
pp. 1-15 ◽  
Author(s):  
Yi Huang ◽  
Bin Xia ◽  
Lei Yang
Keyword(s):  
Land Use ◽  
Spatial Distribution ◽  
Emission Intensity ◽  
Carbon Emission ◽  
Agricultural Land ◽  
Industrial Land ◽  
Carbon Emission Intensity ◽  
Land Use Types ◽  
Distribution Structure ◽  
Concentration Degree

This study attempts to discuss the relationship between land use spatial distribution structure and energy-related carbon emission intensity in Guangdong during 1996–2008. We quantized the spatial distribution structure of five land use types including agricultural land, industrial land, residential and commercial land, traffic land, and other land through applying spatial Lorenz curve and Gini coefficient. Then the corresponding energy-related carbon emissions in each type of land were calculated in the study period. Through building the reasonable regression models, we found that the concentration degree of industrial land is negatively correlated with carbon emission intensity in the long term, whereas the concentration degree is positively correlated with carbon emission intensity in agricultural land, residential and commercial land, traffic land, and other land. The results also indicate that land use spatial distribution structure affects carbon emission intensity more intensively than energy efficiency and production efficiency do. These conclusions provide valuable reference to develop comprehensive policies for energy conservation and carbon emission reduction in a new perspective.

scholarly journals USING LANDSAT 8 IMAGE TIME SERIES FOR CROP MAPPING IN A REGION OF CERRADO, BRAZIL

2016 ◽  
Vol XLI-B8 ◽  
pp. 845-850 ◽  
Author(s):  
H. Bendini ◽  
I. D. Sanches ◽  
T. S. Körting ◽  
L. M. G. Fonseca ◽  
A. J. B. Luiz ◽  
...  
Keyword(s):  
Land Use ◽  
Time Series ◽  
Vegetation Index ◽  
Agricultural Land ◽  
Vegetation Indices ◽  
Agricultural Land Use ◽  
Landsat 8 ◽  
Land Use Classification ◽  
Crop Mapping ◽  
Phenological Metrics

The objective of this research is to classify agricultural land use in a region of the Cerrado (Brazilian Savanna) biome using a time series of Enhanced Vegetation Index (EVI) from Landsat 8 OLI. Phenological metrics extracted from EVI time series, a Random Forest algorithm and data mining techniques are used in the process of classification. The area of study is a region in the Cerrado in a region of the municipality of Casa Branca, São Paulo state, Brazil. The results are encouraging and demonstrate the potential of phenological parameters obtained from time series of OLI vegetation indices for agricultural land use classification.

scholarly journals Microbial Community, Biomass and Physico-Chemical Properties of Soil in dry tropics

2020 ◽  
Author(s):  
Weldesemayat Gorems Woldemariam ◽  
Nandita Ghoshal
Keyword(s):  
Land Use ◽  
Microbial Biomass ◽  
Organic Carbon ◽  
Agricultural Land ◽  
Natural Forest ◽  
Degraded Forest ◽  
Microbial Properties ◽  
Physico Chemical ◽  
Land Use Types ◽  
Bamboo Plantation

Abstract Soil physicochemical and microbial properties can be regarded as an important tool to assess soil quality and health. Studying the soil properties under different land use types is great practical significant for land use and soil management regarding soil carbon dynamics and climate change mitigation. However, the changes in land-use types and their effects on soil physicochemical and microbial properties are largely debated and rather unclear. Four different land use types were used to study soil microbial and soil physico-chemical properties. Soil organic carbon and total nitrogen, soil microbial biomass and microbial diversity were determined by micro kjeldahl method, fumigation and extraction method and FAME GC-Ms, respectively. Among all land use pattern the highest water holding capacity (40.06±0.74%), porosity (0.539±0.011%), soil macro-aggregates (64.16±2.64%), organic carbon (0.84±0.054%), total nitrogen (0.123±0.013%), microbial biomass carbon (570.65±35.05 μg/g) and nitrogen (84.21±3.186 μg/g), basal respiration (3.64±0.064μg/g) and b-glucosidase (809.68±39.7μgμg PNP g-1 dry soil h-1) were found to be under natural forest followed by in decreasing order bamboo plantation, degraded forest and agricultural land. Significant differences were observed among the land use types with microbial biomass carbon and B-glucosidase activity. Furthermore, the correlation of analysis showed that microbial biomass, organic carbon, b-glucosidas activity, total nitrogen, moisture content, porosity, water holding capacity, soil macro aggregates were positively correlated to each other and negatively correlated with bulk density, meso and micro soil aggregates at p<0.05. The PLFA analysis showed that microbial community diversity exhibited distinct patterns among land-use types. The conversions of natural forest to other land use type, the amount of PLFA were reduced significantly. The natural forest had high microbial diversity followed by in decreasing order bamboo plantation, degraded forest and agricultural land. Among the organisms G- bacteria and fungi were showed decreasing order from natural forest, bamboo plantation, degraded forest and agricultural land. The reverse was true for G+ bacteria. The result of this study showed that soil physico-chemical and microbial properties were significantly affected by land use types. Thus bamboo based fallow has the potential for improving soil quality and properties in the short term.

scholarly journals Assessing the Impact of land-use types on the groundwater quality: a case study of Mid River Njoro Catchment, Kenya

2020 ◽  
Vol 9 (4) ◽  
Author(s):  
Halake Guyo Rendilicha ◽  
Patrick Home ◽  
James M. Raude ◽  
Charles M. M’Erimba ◽  
Stellamaris Muthoka
Keyword(s):  
Land Use ◽  
Groundwater Quality ◽  
Forest Cover ◽  
Agricultural Land ◽  
Settlement Area ◽  
Mixed Area ◽  
Land Use Types ◽  
The Mean ◽  
The Impact

The study assessed the impact of land-use types on the groundwater quality of the mid River Njoro catchment, Kenya. Groundwater samples were collected from eight boreholes between the period of October 2017 to February 2018 and analyzed for pH, temperature, electrical conductivity, dissolved oxygen, nitrate, ammonium, and total phosphorus. These parameters were used to calculate the Groundwater Quality Index (GQI) value of the study area. The concentration maps (“primary maps I”) were constructed using Kriging interpolation of ArcGIS software from the seven groundwater quality parameters. The “primary maps I” were standardized with the KEBS and WHO standards to the “primary maps II” for ease of integration into a GIS environment. The “primary maps II” were then rated and weighted using a polynomial function to generate “rank maps” before calculating the GQI using spatial analyst tools of ArcGIS software. The land use map was prepared from a high-resolution Google earth satellite imagery of 2015. The mean GQI values for the different land use polygons were calculated and compared using GIS techniques. The GQI ranged from 68.38 to 70.92, indicating a high groundwater quality of mid River Njoro catchment. The major land-use types identified include settlement area, forest cover, agricultural land and mixed area. The agricultural land dominated the study area, followed by settlement area, forest cover and finally mixed area. The mean GQI value in each land use type varied minimally and this could be because of the diffuse nature of the land use types of the study area. Settlement area had low GQI, followed by agricultural land, mixed area and the forest cover had the highest mean GQI value, which corresponds to good quality of groundwater. Even though the variation is insignificant in this particular study, it somehow indicates the adverse effects of different land use on the quality of groundwater.

Sign in / Sign up

Export Citation Format

Share Document