Sentinel-2 cropland mapping using pixel-based and object-based time-weighted dynamic time warping analysis

Belgiu, Mariana; Csillik, Ovidiu

doi:10.1016/j.rse.2017.10.005

Cited by 711 publications

(460 citation statements)

References 60 publications

Supporting

Mentioning

444

Contrasting

Unclassified

Order By: Relevance

“…Random Forest (RF) (Breiman, 2001) is one of the most popular ensemble classifier which produces excellent results for various remote sensing applications (Belgiu, 2018;Pal, 2005). An ensemble classifier consists of multiple classifiers, usually producing better classification results when compared to an individual classifier that is used to build ensemble (Dietterich, 2002).…”

Section: Random Forest Classifiermentioning

confidence: 99%

“…The Agricultural Monitoring Community of Practice of the Group on Earth Observations (GEO), with its Integrated Global Observing Strategy (IGOL), also calls for an operational system in order to monitor the global agriculture using remote sensing (Belgiu, 2018). In literature, there are many studies for Land Use Land Cover classification as well some of them are dedicated to vegetation mapping used various supervised and unsupervised algorithms in pixel based or object based frameworks (Belgiu, 2018;Chuang, 2016;Nay, 2018;Colkesen, 2017;Li, 2014). A meta-analysis on supervised pixel based techniques for land cover classification performed by Khatami et.al (2016) reveals that inclusion of ancillary data, texture, multi-angle and temporal images gives significant improvement in accuracy of classification.…”

Section: Introductionmentioning

confidence: 99%

“…Sentinel-2 has great potential for spatially-detailed retrievals of phenology (Vrieling, 2018). Belgiu (2018) used time-weighted dynamic time warping (TWDTW) method for crop land mapping on time series Sentinel-2 data by adopting pixel based and object based classification by considering three different study areas and concludes that object-based classification give better results than pixel-based approach. Another study for tea crop mapping has been carried by (Chuang, 2016) using WorldView-2 imagery and machine learning techniques (RF and SVM) and results show that highest overall accuracy is achieved in OBIA.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Exploring Capabilities of Sentinel-2 for Vegetation Mapping Using Random Forest

Saini

Ghosh

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Accurate vegetation mapping is essential for monitoring crop and sustainable agricultural practice. This study aims to explore the capabilities of Sentinel-2 data over Landsat-8 Operational Land Imager (OLI) data for vegetation mapping. Two combination of Sentinel-2 dataset have been considered, first combination is 4-band dataset at 10m resolution which consists of NIR, R, G and B bands, while second combination is generated by stacking 4 bands having 10m resolution along with other six sharpened bands using Gram-Schmidt algorithm. For Landsat-8 OLI dataset, six multispectral bands have been pan-sharpened to have a spatial resolution of 15m using Gram-Schmidt algorithm. Random Forest (RF) and Maximum Likelihood classifier (MLC) have been selected for classification of images. It is found that, overall accuracy achieved by RF for 4-band, 10-band dataset of Sentinel-2 and Landsat-8 OLI are 88.38%, 90.05% and 86.68% respectively. While, MLC give an overall accuracy of 85.12%, 87.14% and 83.56% for 4-band, 10-band Sentinel and Landsat-8 OLI respectively. Results shown that 10-band Sentinel-2 dataset gives highest accuracy and shows a rise of 3.37% for RF and 3.58% for MLC compared to Landsat-8 OLI. However, all the classes show significant improvement in accuracy but a major rise in accuracy is observed for Sugarcane, Wheat and Fodder for Sentinel 10-band imagery. This study substantiates the fact that Sentinel-2 data can be utilized for mapping of vegetation with a good degree of accuracy when compared to Landsat-8 OLI specifically when objective is to map a sub class of vegetation.

show abstract

Section: Random Forest Classifiermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Exploring Capabilities of Sentinel-2 for Vegetation Mapping Using Random Forest

Saini

Ghosh

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…A number of approaches have been devised to extract field boundaries from satellite imagery, which provides regular and global coverage of cropping areas at high resolution, but these methods also tend to over-segment fields with a high internal variability and under-segment small adjacent fields (Belgiu and Csillik, 2018). Some of these adverse effects might be mitigated by purposefully over-segmenting images and deciding whether adjacent objects should be merged with machine learning (see Garcia-Pedrero et al, 2017, for instance).…”

Section: Introductionmentioning

confidence: 99%

Extracting field boundaries from satellite imagery with a convolutional neural network to enable smart farming at scale

Waldner¹,

Diakogiannis²

2020

Preprint

View full text Add to dashboard Cite

Applications of digital agricultural services often require either farmers or their advisers to provide digital records of their field boundaries. Automatic extraction of field boundaries from satellite imagery would reduce the reliance on manual input of these records which is time consuming and error-prone, and would underpin the provision of remote products and services. The lack of current field boundary data sets seems to indicate low uptake of existing methods, presumably because of expensive image preprocessing requirements and local, often arbitrary, tuning. In this paper, we address the problem of field boundary extraction from satellite images as a multi-task semantic segmentation problem. We used ResUNet-a, a deep convolutional neural network with a fully connected UNet backbone that features dilated convolutions and conditioned inference, to assign three labels to each pixel: 1) the probability of belonging to a field; 2) the probability of being part of a boundary; and 3) the distance to the closest boundary. These labels can then be combined to obtain closed field boundaries. Using a single composite image from Sentinel-2, the model was highly accurate in mapping field extent, field boundaries, and, consequently, individual fields. Replacing the monthly composite with a single-date image close to the compositing period only marginally decreased accuracy. We then showed in a series of experiments that our model generalised well across resolutions, sensors, space and time without recalibration. Building consensus by averaging model predictions from at least four images acquired across the season is the key to coping with the temporal variations of accuracy. By minimising image preprocessing requirements and replacing local arbitrary decisions by datadriven ones, our approach is expected to facilitate the extraction of individual crop fields at scale.

show abstract

“…This lack of spatial information may result in a less effective planning of valuable agriculture resources within a district. Advances in satellite remote sensing has proven to be an effective tool for crop mapping and made it possible to estimate crops area at coarser resolution of one kilometer to finer resolution within few meters (Zhou et al, 2013;Jiao et al, 2014;Qin et al, 2015;Belgiu and Csillik, 2018). For small study areas, high-resolution images are used to identify different crops based on supervised classification approach using field signatures (Blaschke, 2010;Yang et al, 2011;Huang et al, 2017).…”

Section: Introductionmentioning

confidence: 99%

Evaluation of Different Phenological Information to Map Crop Rotation in Complex Irrigated Indus Basin

Ismaeel

Zhou

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Accurate information of crop rotation in large basin is essential for policy decisions on land, water and nutrient resources around the world. Crop area estimation using low spatial resolution remote sensing data is challenging in a large heterogeneous basin having more than one cropping seasons. This study aims to evaluate the accuracy of two phenological datasets individually and in combined form to map crop rotations in complex irrigated Indus basin without image segmentation. Phenology information derived from Normalized Difference Vegetation Index (NDVI) and Leaf Area Index (LAI) of Moderate Resolution Imaging Spectroradiometer (MODIS) sensor, having 8-day temporal and 1000 m spatial resolution, was used in the analysis. An unsupervised (temporal space clustering) to supervised (area knowledge and phenology behavior) classification approach was adopted to identify 13 crop rotations. Estimated crop area was compared with reported area collected by field census. Results reveal that combined dataset (NDVI*LAI) performs better in mapping wheat-rice, wheat-cotton and wheat-fodder rotation by attaining root mean square error (RMSE) of 34.55, 16.84, 20.58 and mean absolute percentage error (MAPE) of 24.56%, 36.82%, 30.21% for wheat, rice and cotton crop respectively. For sugarcane crop mapping, LAI produce good results by achieving RMSE of 8.60 and MAPE of 34.58%, as compared to NDVI (10.08, 40.53%) and NDVI*LAI (10.83, 39.45%). The availability of major crop rotation statistics provides insight to develop better strategies for land, water and nutrient accounting frameworks to improve agriculture productivity.

show abstract

Sentinel-2 cropland mapping using pixel-based and object-based time-weighted dynamic time warping analysis

Cited by 711 publications

References 60 publications

Exploring Capabilities of Sentinel-2 for Vegetation Mapping Using Random Forest

Exploring Capabilities of Sentinel-2 for Vegetation Mapping Using Random Forest

Extracting field boundaries from satellite imagery with a convolutional neural network to enable smart farming at scale

Evaluation of Different Phenological Information to Map Crop Rotation in Complex Irrigated Indus Basin

Contact Info

Product

Resources

About