Spatial-Temporal Conditional Random Fields Crop Classification From Terrasar-X Images

Kenduiywo, Benson Kipkemboi; Bargiel, Damian; Soergel, Uwe

doi:10.5194/isprsannals-ii-3-w4-79-2015

Cited by 10 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other studies proposed to introduce a temporal structure using Hidden Markov Chains in a classification pipeline but aimed at modeling phenology instead of crop rotations (Aurdal, Huseby, Eikvil et al 2005;Leite, Feitosa, Formaggio et al 2011;Siachalou, Mallinis, Tsakiri-Strati 2015). Kenduiywoa, Bargiel, and Soergel (2015) modeled phenology information into a conditional random field (CRF), but the classification was performed at different dates through the year. The CRFs were used for classifying land cover classes and crop types on mono-temporal Landsat data (Roscher, Waske, Förstner 2017).…”

Section: Crop Rotation Integrationmentioning

confidence: 99%

Improved Crop Classification with Rotation Knowledge using Sentinel-1 and -2 Time Series

Giordano¹,

Bailly²,

Landrieu³

et al. 2020

photogramm eng remote sensing

View full text Add to dashboard Cite

Leveraging the recent availability of accurate, frequent, and multimodal (radar and optical) Sentinel-1 and -2 acquisitions, this paper investigates the automation of land parcel identi- fication system (LPIS ) crop type classification. Our approach allows for the automatic integration of temporal knowledge, i.e., crop rotations using existing parcel-based land cover databases and multi-modal Sentinel-1 and -2 time series. The temporal evolution of crop types was modeled with a linear- chain conditional random field, trained with time series of multi-modal (radar and optical) satellite acquisitions and associated LPIS. Our model was tested on two study areas in France (≥ 1250 km2) which show different crop types, various parcel sizes, and agricultural practices: . the Seine et Marne and the Alpes de Haute-Provence classified accordingly to a fine national 25-class nomenclature. We first trained a Random Forest classifier without temporal structure to achieve 89.0% overall accuracy in Seine et Marne (10 classes) and 73% in Alpes de Haute-Provence (14 classes). We then demonstrated experimentally that taking into account the temporal structure of crop rotation with our model resulted in an increase of 3% to +5% in accuracy. This increase was especially important (+12%) for classes which were poorly classified without using the temporal structure. A stark posi- tive impact was also demonstrated on permanent crops, while it was fairly limited or even detrimental for annual crops.

show abstract

Section: Crop Rotation Integrationmentioning

confidence: 99%

Improved Crop Classification with Rotation Knowledge using Sentinel-1 and -2 Time Series

Giordano¹,

Bailly²,

Landrieu³

et al. 2020

photogramm eng remote sensing

View full text Add to dashboard Cite

show abstract

“…Recent studies [20], [21] have incorporated knowledgebased crop phenological information characterizing polarimetric parameters for multi-temporal classification. An improvement in the classification accuracy was reported while using crop phenological information in the form of a spatialtemporal sequence pattern (phenological sequence pattern, PSP) [22].…”

Section: Introductionmentioning

confidence: 99%

A Novel Phenology Based Feature Subset Selection Technique Using Random Forest for Multitemporal PolSAR Crop Classification

Hariharan

Mandal

Tirodkar

et al. 2018

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Feature selection techniques intent to select a subset of features that minimizes redundancy and maximizes relevancy for classification problems in machine learning. Standard methods for feature selection in machine learning seldom take into account the domain knowledge associated with the data. Multi-temporal crop classification studies with full-polarimetric Synthetic Aperture Radar (PolSAR) data ought to consider the changes in the scattering mechanisms with their phenological growth stages. Hence, it is desirable to incorporate these changes while determining a feature subset for classification. In this study, a Random Forest (RF) based feature selection technique is proposed which takes into account the changes in the physical scattering mechanism with crop phenological stages for multitemporal PolSAR classification. The partial probability plot, which is an attribute of RF, provides information about the marginal effect of a polarimetric parameter on the desired crop class. Moreover, it is used to identify the specific range of a parameter where the probability of the presence of a particular crop class is high. The proposed technique identifies features which change significantly with crop phenology. The selected features are the ones whose ranges show maximum separation amongst crop classes. Additionally, the feature subset is refined by eliminating correlated features. The E-SAR L-band dataset of the AgriSAR-2006 campaign over the Demmin test site in Germany is used in this study. The classification accuracy using the novel feature selection technique is 99.12%. This is nominally better than using the features obtained from a standard feature selection method used in RF such as Mean Decrease Gini (98.73%) and Mean Decrease Accuracy (98.68%) which do not take into account the information based on crop phenology.

show abstract

“…Approaches that take the temporal dependencies into account usually model temporal interaction by class transition matrices that can be determined by an expert (Hoberg et al, 2010) (Hoberg et al, 2011) empirically from existing data sources, * Corresponding author or computed statistically (Leite et al, 2011) (Kenduiywo et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

A Method to Estimate Temporal Interaction in a Conditional Random Field Based Approach for Crop Recognition

Diaz¹,

Feitosa²,

Sanches³

et al. 2016

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

View full text Add to dashboard Cite

ABSTRACT:This paper presents a method to estimate the temporal interaction in a Conditional Random Field (CRF) based approach for crop recognition from multitemporal remote sensing image sequences. This approach models the phenology of different crop types as a CRF. Interaction potentials are assumed to depend only on the class labels of an image site at two consecutive epochs. In the proposed method, the estimation of temporal interaction parameters is considered as an optimization problem, whose goal is to find the transition matrix that maximizes the CRF performance, upon a set of labelled data. The objective functions underlying the optimization procedure can be formulated in terms of different accuracy metrics, such as overall and average class accuracy per crop or phenological stages. To validate the proposed approach, experiments were carried out upon a dataset consisting of 12 co-registered LANDSAT images of a region in southeast of Brazil. Pattern Search was used as the optimization algorithm. The experimental results demonstrated that the proposed method was able to substantially outperform estimates related to joint or conditional class transition probabilities, which rely on training samples.

show abstract

Spatial-Temporal Conditional Random Fields Crop Classification From Terrasar-X Images

Cited by 10 publications

References 30 publications

Improved Crop Classification with Rotation Knowledge using Sentinel-1 and -2 Time Series

Improved Crop Classification with Rotation Knowledge using Sentinel-1 and -2 Time Series

A Novel Phenology Based Feature Subset Selection Technique Using Random Forest for Multitemporal PolSAR Crop Classification

A Method to Estimate Temporal Interaction in a Conditional Random Field Based Approach for Crop Recognition

Contact Info

Product

Resources

About