An Evaluation of Different Training Sample Allocation Schemes for Discrete and Continuous Land Cover Classification Using Decision Tree-Based Algorithms

Colditz, René R.

doi:10.3390/rs70809655

Cited by 148 publications

(92 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Consequently, "irrigated rice" was classified with a high f1-score of 0.82, allowing for classification of this key crop in terms of food security, with a high level of confidence. All of these results should, however, be interpreted with in mind the fact that the training set represented only 0.038% of the total number of objects in the study area (more the 6 million), whereas studies like [48] recommend a training set of 0.25% of the whole study area, meaning more than 15,000 training samples in our case. Such maps could be validated according to the method provided by [49] for area-based and location-based validation of the classified images using OBIA, in order to spatially identify where there is error or uncertainty.…”

Section: Relevance Of the Approachmentioning

confidence: 99%

A Combined Random Forest and OBIA Classification Scheme for Mapping Smallholder Agriculture at Different Nomenclature Levels Using Multisource Data (Simulated Sentinel-2 Time Series, VHRS and DEM)

et al. 2017

View full text Add to dashboard Cite

Sentinel-2 images are expected to improve global crop monitoring even in challenging tropical small agricultural systems that are characterized by high intra-and inter-field spatial variability and where satellite observations are disturbed by the presence of clouds. To overcome these constraints, we analyzed and optimized the performance of a combined Random Forest (RF) classifier/object-based approach and applied it to multisource satellite data to produce land use maps of a smallholder agricultural zone in Madagascar at five different nomenclature levels. The RF classifier was first optimized by reducing the number of input variables. Experiments were then carried out to (i) test cropland masking prior to the classification of more detailed nomenclature levels, (ii) analyze the importance of each data source (a high spatial resolution (HSR) time series, a very high spatial resolution (VHSR) coverage and a digital elevation model (DEM)) and data type (spectral, textural or other), and (iii) quantify their contributions to classification accuracy levels. The results show that RF classifier optimization allowed for a reduction in the number of variables by 1.5-to 6-fold (depending on the classification level) and thus a reduction in the data processing time. Classification results were improved via the hierarchical approach at all classification levels, achieving an overall accuracy of 91.7% and 64.4% for the cropland and crop subclass levels, respectively. Spectral variables derived from an HSR time series were shown to be the most discriminating, with a better score for spectral indices over the reflectances. VHSR data were only found to be essential when implementing the segmentation of the area into objects and not for the spectral or textural features they can provide during classification.

show abstract

Section: Relevance Of the Approachmentioning

confidence: 99%

A Combined Random Forest and OBIA Classification Scheme for Mapping Smallholder Agriculture at Different Nomenclature Levels Using Multisource Data (Simulated Sentinel-2 Time Series, VHRS and DEM)

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Sample selection was class balanced, as recommended by Belgiu and Drăguţ [17], through matching the object number to the least represented class, which was the weed class in most of the scenarios, e.g., in the C2-26 field, 23% of the data corresponded to bare soil objects, 28% to crop, 28% to shadow and 22% to weeds Thus, weed abundance in each field determined the percentage of objects selected for every class, as well as the full data set composition (second block in Table 2). The full data represented more than 0.25% of the total field area in all scenarios, as the minimum value recommended for training the RF algorithm [46]. For example, the full training data reached 4.2% in S1-16 at 30 m flight altitude, and the 6.8% in S2-16 for images taken at 60 m. * selected training objects: percentage of objects selected for every class in the total training set data for the image.…”

Section: Automatic Rf Training Set Selection and Classificationmentioning

confidence: 99%

An Automatic Random Forest-OBIA Algorithm for Early Weed Mapping between and within Crop Rows Using UAV Imagery

et al. 2018

View full text Add to dashboard Cite

Accurate and timely detection of weeds between and within crop rows in the early growth stage is considered one of the main challenges in site-specific weed management (SSWM). In this context, a robust and innovative automatic object-based image analysis (OBIA) algorithm was developed on Unmanned Aerial Vehicle (UAV) images to design early post-emergence prescription maps. This novel algorithm makes the major contribution. The OBIA algorithm combined Digital Surface Models (DSMs), orthomosaics and machine learning techniques (Random Forest, RF). OBIA-based plant heights were accurately estimated and used as a feature in the automatic sample selection by the RF classifier; this was the second research contribution. RF randomly selected a class balanced training set, obtained the optimum features values and classified the image, requiring no manual training, making this procedure time-efficient and more accurate, since it removes errors due to a subjective manual task. The ability to discriminate weeds was significantly affected by the imagery spatial resolution and weed density, making the use of higher spatial resolution images more suitable. Finally, prescription maps for in-season post-emergence SSWM were created based on the weed maps-the third research contribution-which could help farmers in decision-making to optimize crop management by rationalization of the herbicide application. The short time involved in the process (image capture and analysis) would allow timely weed control during critical periods, crucial for preventing yield loss.

show abstract

“…At first glance, the large number of over 185,000 digitized training polygons might seem to be overdrawn, but it was required for obtaining representative samples for imagery from different acquisition times and years and was necessary because NFI data were only used as an independent data set for validation. Due to the sensitivity of RF classification to the sampling design and imbalanced training samples [51,52], the training samples for the tree type map had to be balanced and representative of the two classes, i.e., area-proportional, and large enough to accommodate the large number of data dimensions. In fact, reducing the number of samples to a certain degree would only slightly lower the model accuracies [7] but would have a greater (negative) impact on the prediction of the tree types.…”

Section: General Aspects Of the Tree Type Mapping Approachmentioning

confidence: 99%

Wall-to-Wall Tree Type Mapping from Countrywide Airborne Remote Sensing Surveys

2017

View full text Add to dashboard Cite

Abstract:Although wall-to-wall, accurate, and up-to-date forest composition maps at the stand level are a fundamental input for many applications, ranging from global environmental issues to local forest management planning, countrywide mapping approaches on the tree type level remain rare. This paper presents and validates an innovative remote sensing based approach for a countrywide mapping of broadleaved and coniferous trees in Switzerland with a spatial resolution of 3 m. The classification approach incorporates a random forest classifier, explanatory variables from multispectral aerial imagery and a Digital Terrain Model (DTM) from Airborne Laser Scanning (ALS) data, digitized training polygons and independent validation data from the National Forest Inventory (NFI). The methodological workflow was optimized for an area of 41,285 km 2 that is characterized by temperate forests within a complex topography. Whereas high model overall accuracies (0.99) and kappa (0.98) were achieved, the comparison of the tree type map with independent NFI data revealed significant deviations that are related to underestimations of broadleaved trees (median of −3.17%). Constraints of the tree type mapping approach are mostly related to the acquisition date and time of the imagery and the topographic (negative) effects on the prediction. A comparison with the most recent High Resolution Layers (HRL) forest 2012 from the European Environmental Agency revealed that the tree type map is superior regarding spatial resolution, level of detail and accuracy. The high-quality map achieved with the approach presented here is of great value for optimizing forest management and planning activities and is also an important information source for applications outside the forestry sector.

show abstract

An Evaluation of Different Training Sample Allocation Schemes for Discrete and Continuous Land Cover Classification Using Decision Tree-Based Algorithms

Cited by 148 publications

References 52 publications

A Combined Random Forest and OBIA Classification Scheme for Mapping Smallholder Agriculture at Different Nomenclature Levels Using Multisource Data (Simulated Sentinel-2 Time Series, VHRS and DEM)

A Combined Random Forest and OBIA Classification Scheme for Mapping Smallholder Agriculture at Different Nomenclature Levels Using Multisource Data (Simulated Sentinel-2 Time Series, VHRS and DEM)

An Automatic Random Forest-OBIA Algorithm for Early Weed Mapping between and within Crop Rows Using UAV Imagery

Wall-to-Wall Tree Type Mapping from Countrywide Airborne Remote Sensing Surveys

Contact Info

Product

Resources

About