Mapping crop cover using multi-temporal Landsat 8 OLI imagery

Crop classification maps are useful for estimating amounts of crops harvested, which could help address challenges in food security. Remote-sensing techniques are useful tools for generating crop maps. Optical remote sensing is one of the most attractive options because it offers vegetation indices (VIs) with frequent revisits and has adequate spatial and spectral resolution and some data has been distributed free of charge. However, sufficient consideration has not been given to the potential of VIs calculated from Landsat 8 Operational Land Imager (OLI) data. This article describes the use of Landsat 8 OLI data for the classification of crops in Hokkaido, Japan. In addition to reflectance, VIs calculated from simple formulas that consisted of combinations of two or more reflectance wavebands were evaluated, as well as the six components of the Kauth–Thomas transform. The VIs based on shortwave infrared bands (bands 6 or 7) improved classification accuracy, and using a combination of all derived data from Landsat 8 OLI data resulted in an overall accuracy of 94.5% (allocation disagreement = 4.492 and quantity disagreement = 1.017).     

Object-based delineation of urban tree canopy: assessing change in Oklahoma City, 2006–2013

With a burgeoning global population, the pressures of urbanization are increasingly prevalent. The need to quantify urban greenness remains significant due to environmental impact and its relationship with human well-being. Utilizing 1 m discrete-return airborne lidar-derived digital terrain models (DTMs) and digital surface models (DSMs), aerial imagery, and lidar-imagery fusion, this study assesses vegetation, specifically tree canopy, change within Oklahoma City between 2006 and 2013. Specifically, we (1) identify an accurate object-based image analysis (OBIA) method for the detection of urban vegetation outlines, and (2) apply that method to locate and quantify vegetation change and assess spatial patterns in Oklahoma City between 2006 and 2013. The proposed OBIA approach extracts urban vegetation coverage from aerial imagery and lidar-based models with around 89% accuracy. Regarding vegetation change, Oklahoma City lost 9.69 km² (3.74 mi²) of tree canopy coverage, which accounted for a 2% loss in total greenness.     

Mapping vegetation community types in a highly disturbed landscape: integrating hierarchical object-based image analysis with lidar-derived canopy height data

Focusing on the semi-arid and highly disturbed landscape of San Clemente Island (SCI), California, we test the effectiveness of incorporating a hierarchical object-based image analysis (OBIA) approach with high-spatial resolution imagery and canopy height surfaces derived from light detection and ranging (lidar) data for mapping vegetation communities. The hierarchical approach entailed segmentation and classification of fine-scale patches of vegetation growth forms and bare ground, with shrub species identified, and a coarser-scale segmentation and classification to generate vegetation community maps. Such maps were generated for two areas of interest on SCI, with and without vegetation canopy height data as input, primarily to determine the effectiveness of such structural data on mapping accuracy. Overall accuracy is highest for the vegetation community map derived by integrating airborne visible and near-infrared imagery having very high spatial resolution with the lidar-derived canopy height data. The results demonstrate the utility of the hierarchical OBIA approach for mapping vegetation with very high spatial resolution imagery, and emphasizes the advantage of both multi-scale analysis and digital surface data for accurately mapping vegetation communities within highly disturbed landscapes.     

Selecting patterns and features for between- and within- crop-row weed mapping using UAV-imagery

This paper approaches the problem of weed mapping for precision agriculture, using imagery provided by Unmanned Aerial Vehicles (UAVs) from sunflower and maize crops. Precision agriculture referred to weed control is mainly based on the design of early post-emergence site-specific control treatments according to weed coverage, where one of the most important challenges is the spectral similarity of crop and weed pixels in early growth stages. Our work tackles this problem in the context of object-based image analysis (OBIA) by means of supervised machine learning methods combined with pattern and feature selection techniques, devising a strategy for alleviating the user intervention in the system while not compromising the accuracy. This work firstly proposes a method for choosing a set of training patterns via clustering techniques so as to consider a representative set of the whole field data spectrum for the classification method. Furthermore, a feature selection method is used to obtain the best discriminating features from a set of several statistics and measures of different nature. Results from this research show that the proposed method for pattern selection is suitable and leads to the construction of robust sets of data. The exploitation of different statistical, spatial and texture metrics represents a new avenue with huge potential for between and within crop-row weed mapping via UAV-imagery and shows good synergy when complemented with OBIA. Finally, there are some measures (specially those linked to vegetation indexes) that are of great influence for weed mapping in both sunflower and maize crops.     

Mapping and discrimination of soya bean and corn crops using spectro-temporal profiles of vegetation indices

The use of remote-sensing technology has been studied as a way to make the monitoring of agricultural crops more efficient, dynamic, and reliable. The use of data from the Moderate Resolution Imaging Spectroradiometer (MODIS) has proved to be an interesting tool regarding the mapping of large areas, however, some challenges still need to be addressed. One of these is the identification of specific types of crops, especially when they have similar phenologies. The purpose of this study was to perform discrimination and mapping of soya bean and corn crops in the state of Paraná, Brazil, for the 2010/2011 and 2011/2012 crop years. A methodology using spectro-temporal profile information of the crops derived from vegetation indices (VIs), the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and the wide dynamic range vegetation index (WDRVI) based on MODIS data was appraised. This method generated a series of maps of the respective crops that were later qualitatively or quantitatively appraised. Some of the maps drawn showed a global accuracy rate above 80% and a kappa coefficient (κ) of over 0.7. The data areas showed an average difference of 6% for the cultivation of soya beans, and 11% for corn when compared to official data. The WDRVI and EVI were similar and showed better performance when compared to the NDVI in the assessments made. The results demonstrate that the soya bean crop was better mapped compared to corn, particularly in terms of the size of the crop area. The use of spectro-temporal profiles of the VIs assisted in obtaining important information, enabling better identification of crops from regional scale mapping using the MODIS data.     

Distinguishing wetland vegetation and channel features with object-based image segmentation

Mapping landscape features within wetlands using remote-sensing imagery is a persistent challenge due to the fine scale of wetland pattern variation and the low spectral contrast among plant species. Object-based image analysis (OBIA) is a promising approach for distinguishing wetland features, but systematic guidance for this use of OBIA is not presently available. A sensitivity analysis was tested using OBIA to distinguish vegetation zones, vegetation patches, and surface water channels in two intertidal salt marshes in southern San Francisco Bay. Optimal imagery sources and OBIA segmentation settings were determined from 348 sensitivity tests using the eCognition multiresolution segmentation algorithm. The optimal high-resolution (≤1 m) imagery choices were colour infrared (CIR) imagery to distinguish vegetation zones, CIR or red, green, blue (RGB) imagery to distinguish vegetation patches depending on species and season, and RGB imagery to distinguish surface water channels. High-resolution (1 m) lidar data did not help distinguish small surface water channels or other features. Optimal segmentation varied according to segmentation setting choices. Small vegetation patches and narrow channels were more recognizable using small scale parameter settings and coarse vegetation zones using larger scale parameter settings. The scale parameter served as a de facto lower bound to median segmented object size. Object smoothness/compactness weight settings had little effect. Wetland features were more recognizable using high colour/low shape weight settings. However, an experiment on a synthetic non-wetland image demonstrated that, colour information notwithstanding, segmentation results are still strongly affected by the selected image resolution, OBIA settings, and shape of the analysis region. Future wetland OBIA studies may benefit from strategically making imagery and segmentation setting choices based on these results; such systemization of future wetland OBIA approaches may also enhance study comparability.     

Airborne multi-spectral monitoring of agricultural crop status: effect of time of year,crop type and crop condition parameter

Airborne optical multispectral imagery was acquired in conjunction with contemporaneous ground-based measurements of various crops (leaf area index, canopy temperature, plant height) at a test site in southern Alberta, Canada. Data were acquired on three occasions in July 1994 for a variety of crops and irrigation practices. A large number of crop condition-spectral relations were examined to determine whether the imagery could be used to measure the various crop condition parameters. It was found that a number of statistically significant correlations exist between the imagery and the crop condition parameters and that these correlations vary as a function of crop type, time of year, and crop condition parameter. The results suggest that in many cases, multi-spectral optical imagery can be used to monitor variations in crop condition parameters across the growing season for a variety of crop types.     

Creating a basic customizable framework for crop detection using Landsat imagery

Remotely sensed crop identification is essential for countries whose economic vitality is closely tied to agriculture, such as Uruguay. It has been shown that using Normalized Difference Vegetation Index (NDVI) can sometimes produce spurious results when classifying land cover in certain environments. Furthermore, many current crop identification tools use NDVI in order to study and identify crop land-cover for classification techniques. In this study, we present the basic framework for a semi-automated crop identification methodology, which uses a time series analysis to identify soil and vegetation patterns for various crop-cycle scenarios by using the pixel Hue values for land cover identification, at high (30 m) spatial resolution. This is accomplished by converting the Red–Green–Blue (RGB) colour space of a shortwave infrared (SWIR), near-infrared, and red channel composite images, into a Hue-Saturation-Value colour space, then extracting the Hue pixel values that correspond to soil and vegetation over a series of images. We then combine the soil and vegetation pixels in order to create a ‘time series’ to identify which pixels match different crop-cycle scenarios and isolate them. The shapes are then further isolated to only include those that fit a specific shape area (>20 ha), in order to eliminate spurious results. Our results show an 80% accuracy score between the crop identification methodology and a proposed crop plan over the years 2013–2014 and probabilities of detection of 0.76, 0.89, and 0.88 for the seasons of 2009–2010, 2010–2011, and 2011–2012 respectively, when compared to verified partial crop location maps. The proposed crop plan and the partial crop location maps were provided to us by the Instituto Nacional de Investigación Agropecuaria (INIA) in Uruguay. We also quantitatively investigated the shortcomings of the crop identification methodology, which mostly came from cloud cover and low temporal resolution of the images.     

Object-based satellite image analysis applied for crop residue estimating using Landsat OLI imagery

ABSTRACT

The main objective of this study is to apply an object-based image analysis (OBIA) approach to satellite image processing and determining crop residue cover (CRC) and tillage intensity. To achieve this goal, we collected ground truth data using line-transect method from 35 plots of farmlands with an area of 528 ha. Accordingly, Landsat Operational Land Imager (OLI) satellite image together with global positioning system (GPS)-based survey data set were considered for applying the OBIA methods and deriving CRC. To process the data, object-based image processing steps including segmentation and classification were applied to develop intelligent objects and establish classification using spectral and spatial characteristics of CRC. We developed three categories of rule sets including mean indices, tillage indices, and grey-level co-occurrence matrix (GLCM) texture features using the OBIA algorithms and assign class method. Results were validated against of ground control data set and were collected by GPS in field survey. Results of this study indicated that the brightness, normalised difference tillage index, and GLCM texture feature mean performed out as effective techniques. Overall accuracy and kappa coefficient (κ) were computed to be about 0.91 and 0.86; 0.93 and 0.90; 0.60 and 0.35, respectively, for the above-mentioned indices. The foregoing discussion has attempted to demonstrate that the remotely sensed data can be effective approach and substitute for ground methods, especially in large areas.     

Estimating crop chlorophyll content with hyperspectral vegetation indices and the hybrid inversion method

A hybrid inversion method was developed to estimate the leaf chlorophyll content (LCC) and canopy chlorophyll content (CCC) of crops. Fifty hyperspectral vegetation indices (VIs), such as the photochemical reflectance index (PRI) and canopy chlorophyll index (CCI), were compared to identify the appropriate VIs for crop LCC and CCC inversion. The hybrid inversion models were then generated from different modelling methods, including the curve-fitting and least squares support vector regression (LS-SVR) and random forest regression (RFR) algorithms, by using simulated Compact High Resolution Imaging Spectrometer (CHRIS) datasets that were generated by a radiative transfer model. Finally, the remote-sensing mapping of a CHRIS image was completed to test the inversion accuracy. The results showed that the remote-sensing mapping of the CHRIS image yielded an accuracy of R² = 0.77 and normalized root mean squared error (NRMSE) = 17.34% for the CCC inversion, and an accuracy of only R² = 0.33 and NRMSE = 26.03% for LCC inversion, which indicates that the remote-sensing technique was more appropriate for obtaining chlorophyll content at the canopy scale (CCC) than at the leaf scale (LCC). The estimated results of various VIs and algorithms suggested that the PRI and CCI were the optimal VIs for LCC and CCC inversion, respectively, and RFR was the optimal method for modelling.     

Estimating life-form cover fractions in California sage scrub communities using multispectral remote sensing

The California sage scrub (CSS) community type in California's Mediterranean-type ecosystems is known for its high biodiversity and is home to a large number of rare, threatened, and endangered species. Because of extensive urban development in the past fifty years, this ecologically significant community type is highly degraded and fragmented. To conserve endangered CSS communities, monitoring internal conditions of communities is as crucial as monitoring distributions of the community type in the region. Vegetation type mapping and field sampling of individual plants provide ecologically meaningful information about CSS communities such as spatial distribution and species compositions, respectively. However, both approaches only provide spatially comprehensive information but no information about internal conditions or vice versa. Therefore, there is a need for monitoring variables which fill the information gap between vegetation type maps and field-based data. A number of field-based studies indicate that life-form fractional cover is an effective indicator of CSS community health and habitat quality for CSS-obligated species. This study investigates the effectiveness of remote sensing approaches for estimating fractional cover of true shrub, subshrub, herb, and bare ground in CSS communities of southern California. Combinations of four types of multispectral imagery ranging from 0.15 m resolution scanned color infrared aerial photography to 10 m resolution SPOT 5 multispectral imagery and three image processing models - per-pixel, object-based, and spectral mixture models - were tested.An object-based image analysis (OBIA) routine consistently yielded higher accuracy than other image processing methods for estimating all cover types. Life-form cover was reliably predicted, with error magnitudes as low as 2%. Subshrub and herb cover types required finer spatial resolution imagery for more accurate predictions than true shrub and bare ground types. Positioning of sampling grids had a substantial impact on the reliability of accuracy assessment, particularly for cover estimates predicted using multiple endmember spectral mixture analysis (MESMA) applied to SPOT imagery. Of the approaches tested in this study, OBIA using pansharpened QuickBird imagery is one of the most promising approaches because of its high accuracy and processing efficiency and should be tested for more heterogeneous CSS landscapes. MESMA applied to SPOT imagery should also be examined for effectiveness in estimating factional cover over more extensive habitat areas because of its low data cost and potential for conducting retrospective studies of vegetation community conditions.     

Estimation of the fractional canopy cover of pecan orchards using Landsat 5 satellite data,aerial imagery,and orchard floor photographs

Pecan orchards are the largest agricultural water consumer in the lower part of the Mesilla Valley, NM, USA. Knowledge of fractional canopy (FC) cover allows better crop water use assessment and orchard management. FC can be estimated from vegetation indices (VIs), such as the normalized difference vegetation index (NDVI), the soil-adjusted vegetation index (SAVI), the simple ratio (SR), and the triangular vegetation index (TVI), using satellite imagery. The main objective of this research is to develop an approach to determine the FC from a simultaneous imagery campaign consisting of aerial imagery, orchard floor photographs, and satellite images. All the required data were collected based on satellite overpass times at three different times during the initial part of the growing season to enhance the quality of data and reduce errors. The data were processed using the software package Environment for Visualizing Images (ENVI^® 4.6.1; ITT Research Systems Inc.). The orchard floor digital photographs were used as a ground truth data set that gave a good correlation to the aerial photography. The aerial images were then used to determine the relationship between the FC and the VIs using these ‘corrected FCs’. The results showed significant correlation between NDVI and FC (R ² = 0.80; p < 0.0001). Likewise, the calculated SR not only showed good correlation to the FCs but also verified the calculated NDVI. The results indicated that the methodology of this research can be applied to other tree crops as an aid in estimating the FC.     

Mapping vegetation morphology types in a dry savanna ecosystem: integrating hierarchical object-based image analysis with Random Forest

Savanna ecosystems are geographically extensive and both ecologically and economically important, and require monitoring over large spatial extents. Remote-sensing-based characterization of vegetation properties in savannas is methodologically challenging, mainly due to high structural and functional heterogeneity. Recent advances in object-based image analysis (OBIA) and machine learning algorithms offer new opportunities to address these challenges. Focusing on the semi-arid savanna ecosystem in the central Kalahari, this study examined the suitability of a hierarchical OBIA approach combined with in situ data and an ensemble classification technique for mapping vegetation morphology types at landscape scale. A stack of Landsat TM imagery, NDVI, and topographic variables was segmented with six different scale factors resulting in a hierarchical network of image objects. Sample objects for each vegetation morphology class were selected at each segmentation scale and classification was performed using optimal features consisting of spectral and textural features. Overall and class-specific classification accuracies were compared across the six scales to examine the influence of segmentation scale on each. Results suggest that the highest overall classification accuracy (i.e. 85.59%) was observed not at the finest segmentation scale, but at coarse segmentation. Additionally, individual vegetation morphology classes differed in the segmentation scale at which they achieved highest classification accuracy, reflecting their unique ecology and physiognomic composition. While classes with high vegetation density/height attained higher accuracy at fine segmentation scale, those with lower vegetation density/height reached higher classification accuracy at coarse segmentation scales. Contrarily, for pans and bare areas, accuracy was relatively unaffected by changing segmentation scale. Variable importance plots suggested that spectral features were the most important, followed by textural variables. These results show the utility of the OBIA approach and emphasize the requirement of multi-scale analysis for accurately characterizing savanna systems.     

A comparison of MODIS 250-m EVI and NDVI data for crop mapping: a case study for southwest Kansas

Multi-temporal vegetation index (VI) data from the Moderate Resolution Imaging Spectroradiometer (MODIS) are becoming widely used for large-area crop classification. Most crop-mapping studies have applied enhanced vegetation index (EVI) data from MODIS instead of the more traditional normalized difference vegetation index (NDVI) data because of atmospheric and background corrections incorporated into EVI's calculation and the index's sensitivity over high biomass areas. However, the actual differences in the classification results using EVI versus NDVI have not been thoroughly explored. This study evaluated time-series MODIS 250-m EVI and NDVI for crop-related land use/land cover (LULC) classification in the US Central Great Plains. EVI- and NDVI-derived maps classifying general crop types, summer crop types and irrigated/non-irrigated crops were produced for southwest Kansas. Qualitative and quantitative assessments were conducted to determine the thematic accuracy of the maps and summarize their classification differences. For the three crop maps, MODIS EVI and NDVI data produced equivalent classification results. High thematic accuracies were achieved with both indices (generally ranging from 85% to 90%) and classified cropping patterns were consistent with those reported for the study area (> 0.95 correlation between the classified and USDA-reported crop areas). Differences in thematic accuracy (< 3% difference), spatially depicted patterns (> 90% pixel-level thematic agreement) and classified crop areas between the series of EVI- and NDVI-derived maps were negligible. Most thematic disagreements were restricted to single pixels or small clumps of pixels in transitional areas between cover types. Analysis of MODIS composite period usage in the classification models also revealed that both VIs performed equally well when periods from a specific growing season phase (green, peak or senescence) were heavily utilized to generate a specific crop map.     

Monitoring Landsat vegetation indices for different crop treatments and soil chemistry

The timing and quantity of fertilizer and herbicide applications in agricultural systems are critical where maximizing vigour and yield is the ultimate goal. While fertilizers are applied to the soil to promote plant growth, herbicides are commonly used to control weeds in order to reduce the weeds’ competition for nutrients. Satellite imagery is frequently used to monitor agricultural activities and vegetation indices (VIs) are widely applied in temporal analysis of crop status. This study considers monitoring Landsat VIs for the period between 5 June and 27 October 2014 in agricultural systems under four different management treatments at the Kellogg Biological Station (KBS), in Michigan, USA. The results show that (1) fine-tuning conventional treatments by intense early herbicide applications in combination with no-tilled soil results in significantly higher VIs during the early growth stage, a more rapid maturity rate, and the highest crop yield; (2) nitrogen uptake from nitrate-based rather than from ammonium-based fertilizers might be more beneficial in terms of crop vigour and yield return; (3) organic treatments, with organic corn and no agricultural chemicals, keep higher VIs longer in the season at the cost of lower yield; and (4) genetically modified (GM) breeds under conventional or reduced-chemical treatments have synchronized early senescence. A positive correlation between VIs during the early growth stage and yield is observed for conventional no-till treatment (coefficient of determination, R² = 0.70). The correlation becomes gradually weaker with each month from late June to October (29 June: R² = 0.70; 16 August: R² = 0.61; 17 September: R² = 0.44; 27 October: R² = 0.01). The analysis of variance (ANOVA)–Tukey–Kramer approach suggests significant differences in VIs between organic and GM corn (treated conventionally or with reduced chemicals) for the preharvest season (27 October 2014). The leave-out-one cross-validation analysis confirms the predictive accuracy of the model (mean square error (MSE) = 0.0014). The rapid evolution of herbicide-resistant weeds requires constant refinement of chemical inputs to agricultural systems, thus making the monitoring of (Landsat) VIs important in the years to come.     

Crop type identification by integration of high-spatial resolution multispectral data with features extracted from coarse-resolution time-series vegetation index data

Crop type identification is the basis of crop acreage estimation and plays a key role in crop production prediction and food security analysis. However, the accuracy of crop type identification using remote-sensing data needs to be improved to support operational agriculture-monitoring tasks. In this paper, a new method integrating high-spatial resolution multispectral data with features extracted from coarse-resolution time-series vegetation index data is proposed to improve crop type identification accuracy in Hungary. Four crop growth features, including peak value, date of peak occurrence, average rate of green-up, and average rate for the senescence period were extracted from time-series Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) profiles and spatially enhanced to 30 m resolution using resolution merge tools based on a multiplicative method to match the spatial resolution of Landsat Thematic Mapper (TM) data. A maximum likelihood classifier (MLC) was used to classify the TM and merged images. Independent validation results indicated that the average overall classification accuracy was improved from 92.38% using TM to 94.67% using the merged images. Based on the classification results using the proposed method, acreages of two major summer crops were estimated and compared to statistical data provided by the United States Department of Agriculture (USDA). The proposed method was able to achieve highly satisfactory crop type identification results.     

Object-based image analysis of optical and radar variables for wetland evaluation

Optical and radar imagery has been shown to be useful for classifying wetland types and surrounding non-wetland classes such as forest and agriculture. Throughout the literature, recommendations have been made that optical and radar image variables together should improve overall and individual class accuracies. object-based image analysis (OBIA) uses multiple data types to segment objects representing land cover entities that are subsequently classified. There are few studies that have utilized optical and polarimetric radar variables together in OBIA to map wetland classes. This research investigated the potential to combine WorldView-2 optical image variables with fully polarimetric Radarsat-2 image variables in OBIA classification of wetland type. With the addition of radar polarimetric variables, classification accuracy improved for the wetland classes of fen, bog, and swamp over the use of optical imagery alone; specifically the addition of Cloude–Pottier (CP) variables of entropy, anisotropy, and alpha angle improved the classification of fen, and the addition of horizontal transmit and horizontal receive (HH) and horizontal transmit and vertical receive (HV) backscatter intensity improved the classification of swamp.     

A comparison of pixel-based decision tree and object-based Support Vector Machine methods for land-cover classification based on aerial images and airborne lidar data

Precisely monitoring land cover/use is crucial for urban environmental assessment and management. Various classification techniques such as pixel-based and object-based approaches have advantages and disadvantages. In this article, based on our experiment data from an unmanned platform carried lidar scanner system and camera, we explored and compared classi?cation accuracies of pixel-based decision tree (DT) and object-based Support Vector Machine (SVM) approaches. Lidar height information can improve classification accuracy based on either object-based SVM or pixel-based DT. From total classification accuracy, object-based SVM was higher than that of pixel-based DT classification, and total accuracy and kappa coefficient of the former were 92.71% and 0.899, respectively. However, pixel-based DT outperformed object-based SVM when classifying small ‘scatter’ tree along roads. Additionally, in order to evaluate the accuracy of pixel-based DT and object-based SVM, we added benchmark data of ISPRS to compare the classification results of two methods. Object-based SVM classification methods by combining aerial imagery with lidar height information can achieve higher classification accuracy. And, accurately extracting tree class of different landscape pattern should select appropriate machine-learning algorithms. Comparison of the results on two methods will provide a reference for selecting a particular classification approaches according to local conditions.     

Urban object extraction using Dempster Shafer feature-based image analysis from worldview-3 satellite imagery

A detailed and up-to-date land use of the urban environment is essentially required in many applications. Very high-resolution (VHR), Multispectral Scanner System (MSS) Worldview-3 (WV-3) satellite imagery provides detailed information on urban characteristics, which should be professionally mined. In this research, WV-3 was processed by machine learning (ML) methods to extract the most accurate urban features. Fuze-Go panchromatic sharpening in conjunction with atmospheric and topographic correction was initially utilized to increase the image quality and colour contrast. Three image analysis approaches including, current pixel-based image analysis (PBIA), object-based image analysis (OBIA) and new feature-based image analysis (FBIA) were implemented on WV-3 image. The k-nearest neighbour (k-NN), Naive Bayes (NB), support vector machine (SVM) classifiers were represented by PBIA, the Decision Tree (DT) classifier was examined as OBIA and the Dempster–Shafer (DS) fusion classifier was manifested for the first time as FBIA. In order to engage DS as FBIA, four types of Belief Masses, namely, Precision, Recall, Overall Accuracy, and kappa coefficient (?) were implemented and compared to assign the most likelihood urban features. All the applied classifiers were also trained on the first site and then tested on another site to examine the transferability. The accuracy, reliability, and computational time of all classifiers were examined by confusion matrix and McNemar assessment. Results show improvements on the detailed urban extraction obtained using the proposed FBIA with 92.2% overall accuracy in compared with PBIA and OBIA. The FBIA result of urban extraction is more consistent when transferred to another study area and consumes much lesser time than OBIA. Also, the precision mass belief measurement achieved highest efficiency regarding receiver operating characteristic (ROC) curve rate.     

Wetland classification using Radarsat-2 SAR quad-polarization and Landsat-8 OLI spectral response data: a case study in the Hudson Bay Lowlands Ecoregion

Circumboreal Canadian bogs and fens distinguished by differences in soils, hydrology, vegetation and morphological features were classified using combinations of Radarsat-2 synthetic aperture radar (SAR) quad-polarization data and Landsat-8 Operational Land Imager (OLI) spectral response patterns. Separate classifications were conducted using a traditional pixel-based maximum likelihood classifer and a machine learning algorithm following an object-based image analysis (OBIA). This study focused on two wetland classes with extensive coverage in the area (bog and fen). In the pixel-based maximum likelihood classification, accuracy increased from approximately 69% user’s accuracy and 79% producer’s accuracy using Radarsat-2 SAR data alone to approximately 80% user’s accuracy and 87% producer’s accuracy using Landsat-8 OLI data alone. Use of the Radarsat-2 SAR and Landsat-8 OLI data following principal components analysis (PCA) data fusion did not result in higher pixel-based maximum likelihood classification accuracy. In the object-based machine learning classification, higher bog and fen class accuracies were obtained when using Radarsat-2 and Landsat OLI data individually compared to the equivalent pixel-based classification. Subsequently, a PCA-data fusion product outperformed the individual bands of the Radarsat-2 and Landsat-8 imagery in object-based classification. Greater than 90% producer’s accuracy was obtained. The margin of error (MOE) was less than 5% in all classifications reported here. Further research will examine alternative data fusion techniques and the addition of Radarsat-2 SAR interferometric digital elevation model (DEM)-based geomorphometrics in object-based classification of different morphological types of bogs and fens.