Object-based Land Cover Classification Research Articles

Graph Neural Network-Based Method of Spatiotemporal Land Cover Mapping Using Satellite Imagery.

Multispectral satellite imagery offers a new perspective for spatial modelling, change detection and land cover classification. The increased demand for accurate classification of geographically diverse regions led to advances in object-based methods. A novel spatiotemporal method is presented for object-based land cover classification of satellite imagery using a Graph Neural Network. This paper introduces innovative representation of sequential satellite images as a directed graph by connecting segmented land region through time. The method's novel modular node classification pipeline utilises the Convolutional Neural Network as a multispectral image feature extraction network, and the Graph Neural Network as a node classification model. To evaluate the performance of the proposed method, we utilised EfficientNetV2-S for feature extraction and the GraphSAGE algorithm with Long Short-Term Memory aggregation for node classification. This innovative application on Sentinel-2 L2A imagery produced complete 4-year intermonthly land cover classification maps for two regions: Graz in Austria, and the region of Portorož, Izola and Koper in Slovenia. The regions were classified with Corine Land Cover classes. In the level 2 classification of the Graz region, the method outperformed the state-of-the-art UNet model, achieving an average F1-score of 0.841 and an accuracy of 0.831, as opposed to UNet's 0.824 and 0.818, respectively. Similarly, the method demonstrated superior performance over UNet in both regions under the level 1 classification, which contains fewer classes. Individual classes have been classified with accuracies up to 99.17%.

LAND COVER CLASSIFICATION PERFORMANCE OF MULTISPECTRAL RTK UAVs

Abstract. Unmanned air vehicle (UAV) became an alternative airborne remote sensing technique, due to providing very high resolution and low cost spatial data and short processing time. Particularly, optical UAVs are frequently utilized in various applications such as mapping, agriculture, and forestry. Especially for precise agriculture purposes, the UAVs were equipped with multispectral cameras which enables to classify land cover easily. In this study, the land cover classification potential of DJI Phantom IV Multispectral, one of the most preferred agricultural UAVs in the world, was investigated using spectral angle mapper, minimum distance and maximum likelihood pixel-based classification techniques and object-based classification. In the investigation, a part of Gebze Technical University (GTU) Northern Campus, includes a large variety of land cover classes, was selected as the study area. The UAV aerial photos were achieved from 70 m flight altitude and processed using structure from motion (SfM)-based image matching software Agisoft Metashape. The pixel-based and object-based land cover classification processes were completed with ENVI and eCognition software respectively. 16 independent land cover classes were classified and the results demonstrated that the accuracies are 73.46% in spectral angle mapper, 75.27% in minimum distance and 93.56% in maximum likelihood pixel-based classification techniques and 90.09% in nearest neighbour object-based classification.

Object-Based on Land Cover Classification on LAPAN-A3 Satellite Imagery Using Tree Algorithm (Case Study: Rote Island)

LAPAN became serious about making a remote sensing satellite on its third-generation satellite. Launched a year after LAPAN-A2, the third-generation satellite, LAPAN-A3, brought LISAT as the main payload. LISAT is a multispectral camera with 4 bands (Red, Green, Blue, NIR) that can be used for land classification, agriculture monitoring, drought monitoring, and land use changing<em>. LAPAN-A3 is the third generation of micro-satellite developed by Satellite Technology Center – LAPAN. This satellite carries a multispectral push-broom sensor that can record the earth's surface at the visible and near-infrared spectrum. </em>This paper aims to determine object-based land cover classification in Rote Island using the LAPAN-A3 satellite image using the tree method algorithm. This classification technique is expected to increase the accuracy of land cover classification. This classification used the LAPAN-A3 satellite imagery of Rote Island. The first process was determined the segmentation with scale parameter 60, shape 0.5, and compactness 0.5. The result shows that OBIA classification on Rote Island, the area of the open land class is 233.67 km², the area of the settlement is 11.57 km², the body of water is 2006.21 km2, the area of low vegetation is 525.93 km², the area of high vegetation is 437.5 km², and there is no data (cloud and cloud shadows) on the LAPAN-A3 image of 45.78 km². The accuracy values obtained were producer 86.67%, KIA 83.02%, Helden 92.86%, Short 86.7%, KIA per class 82.72%, and 85.96%. This object-based classification can meet international and national land cover classification standards, namely at 80%.

UAV-Based Land Cover Classification for Hoverfly (Diptera: Syrphidae) Habitat Condition Assessment: A Case Study on Mt. Stara Planina (Serbia)

Habitat degradation, mostly caused by human impact, is one of the key drivers of biodiversity loss. This is a global problem, causing a decline in the number of pollinators, such as hoverflies. In the process of digitalizing ecological studies in Serbia, remote-sensing-based land cover classification has become a key component for both current and future research. Object-based land cover classification, using machine learning algorithms of very high resolution (VHR) imagery acquired by an unmanned aerial vehicle (UAV) was carried out in three different study sites on Mt. Stara Planina, Eastern Serbia. UAV land cover classified maps with seven land cover classes (trees, shrubs, meadows, road, water, agricultural land, and forest patches) were studied. Moreover, three different classification algorithms—support vector machine (SVM), random forest (RF), and k-NN (k-nearest neighbors)—were compared. This study shows that the random forest classifier performs better with respect to the other classifiers in all three study sites, with overall accuracy values ranging from 0.87 to 0.96. The overall results are robust to changes in labeling ground truth subsets. The obtained UAV land cover classified maps were compared with the Map of the Natural Vegetation of Europe (EPNV) and used to quantify habitat degradation and assess hoverfly species richness. It was concluded that the percentage of habitat degradation is primarily caused by anthropogenic pressure, thus affecting the richness of hoverfly species in the study sites. In order to enable research reproducibility, the datasets used in this study are made available in a public repository.

Object-Based land cover classification of the Vu Gia – Thu Bon river basin on the cloud computing platform

The arrival of cloud computing platform Google Earth Engine (GEE) in 2010 has brought a breakthrough for analysing and processing spatial data. Applying algorithms on this platform has overcome the limitations of commercial software while processing data in building thematic databases, including land cover data. These data are a critical factor for climate change and hydrological models. This study applied Object-based Random Forest (RF) classification in the Google Earth Engine platform to produce land cover data from Landsat 8 data of the Vu Gia - Thu Bon river basin. The classification results showed 7 categories of land cover consisting of artificial forest, natural forest, paddy area, urban area, rural area, bare land, and body water, with an overall accuracy Kappa of 0.70.

Uncertainty Analysis of Object-Based Land-Cover Classification Using Sentinel-2 Time-Series Data

Recently, time-series from optical satellite data have been frequently used in object-based land-cover classification. This poses a significant challenge to object-based image analysis (OBIA) owing to the presence of complex spatio-temporal information in the time-series data. This study evaluates object-based land-cover classification in the northern suburbs of Munich using time-series from optical Sentinel data. Using a random forest classifier as the backbone, experiments were designed to analyze the impact of the segmentation scale, features (including spectral and temporal features), categories, frequency, and acquisition timing of optical satellite images. Based on our analyses, the following findings are reported: (1) Optical Sentinel images acquired over four seasons can make a significant contribution to the classification of agricultural areas, even though this contribution varies between spectral bands for the same period. (2) The use of time-series data alleviates the issue of identifying the “optimal” segmentation scale. The finding of this study can provide a more comprehensive understanding of the effects of classification uncertainty on object-based dense multi-temporal image classification.

Object-based island hierarchical land cover classification using unmanned aerial vehicle multitype data

The unmanned aerial vehicle (UAV) is an emerging technology applied recently in land cover classification, owing to its ability to acquire very high-resolution spatial data, that has provided an effective means for detailed land cover mapping, especially for a small island area. Selecting suitable UAV-acquired data and exploring the combined use of UAV multitype data are of significance for island mapping. Nine classification models were established through a fusion method of visible, multispectral, and light detection and ranging (LIDAR) data acquired by UAVs. A two-level hierarchical land cover classification (level 1 and level 2) of the Donkey Island in China was performed using geographic object-based image analysis with random forest classifier. We investigated the performance of land cover classification models containing different sets of features (spectral, height, intensity, and shape features extracted from UAV data) and evaluated the importance of various features. The results demonstrate that the overall accuracy (OA) of the models generally increase with decreasing detail and the amount of information entering the classification process. The OA achieved range from 82.08% to 92.54% and 74.12% to 85.08% across the nine models for level 1 and level 2, respectively. The best result was achieved with a model combining all features based on multispectral and LIDAR data. Height and intensity information input significantly affect the quality of classification models positively, with height apparently more significant than LIDAR information. When comparing different features, spectral features prominently assist in discriminating land cover classes. The importance of height and intensity features to classification accuracy varies for the classification models, showing greater importance in models based on visible data.

Decision Tree Algorithms for Developing Rulesets for Object-Based Land Cover Classification

Decision tree (DT) algorithms are important non-parametric tools used for land cover classification. While different DTs have been applied to Landsat land cover classification, their individual classification accuracies and performance have not been compared, especially on their effectiveness to produce accurate thresholds for developing rulesets for object-based land cover classification. Here, the focus was on comparing the performance of five DT algorithms: Tree, C5.0, Rpart, Ipred, and Party. These DT algorithms were used to classify ten land cover classes using Landsat 8 images on the Copperbelt Province of Zambia. Classification was done using object-based image analysis (OBIA) through the development of rulesets with thresholds defined by the DTs. The performance of the DT algorithms was assessed based on: (1) DT accuracy through cross-validation; (2) land cover classification accuracy of thematic maps; and (3) other structure properties such as the sizes of the tree diagrams and variable selection abilities. The results indicate that only the rulesets developed from DT algorithms with simple structures and a minimum number of variables produced high land cover classification accuracies (overall accuracy &gt; 88%). Thus, algorithms such as Tree and Rpart produced higher classification results as compared to C5.0 and Party DT algorithms, which involve many variables in classification. This high accuracy has been attributed to the ability to minimize overfitting and the capacity to handle noise in the data during training by the Tree and Rpart DTs. The study produced new insights on the formal selection of DT algorithms for OBIA ruleset development. Therefore, the Tree and Rpart algorithms could be used for developing rulesets because they produce high land cover classification accuracies and have simple structures. As an avenue of future studies, the performance of DT algorithms can be compared with contemporary machine-learning classifiers (e.g., Random Forest and Support Vector Machine).

Dimensionality Reduction and Feature Selection for Object-Based Land Cover Classification based on Sentinel-1 and Sentinel-2 Time Series Using Google Earth Engine

Mapping Earth’s surface and its rapid changes with remotely sensed data is a crucial task to understand the impact of an increasingly urban world population on the environment. However, the impressive amount of available Earth observation data is only marginally exploited in common classifications. In this study, we use the computational power of Google Earth Engine and Google Cloud Platform to generate an oversized feature set in which we explore feature importance and analyze the influence of dimensionality reduction methods to object-based land cover classification with Support Vector Machines. We propose a methodology to extract the most relevant features and optimize an SVM classifier hyperparameters to achieve higher classification accuracy. The proposed approach is evaluated in two different urban study areas of Stockholm and Beijing. Despite different training set sizes in the two study sites, the averaged feature importance ranking showed similar results for the top-ranking features. In particular, Sentinel-2 NDVI, NDWI, and Sentinel-1 VV temporal means are the highest ranked features and the experiment results strongly indicated that the fusion of these features improved the separability between urban land cover and land use classes. Overall classification accuracies of 94% and 93% were achieved in Stockholm and Beijing study sites, respectively. The test demonstrated the viability of the methodology in a cloud-computing environment to incorporate dimensionality reduction as a key step in the land cover classification process, which we consider essential for the exploitation of the growing Earth observation big data. To encourage further research and development of reliable workflows, we share our datasets and publish the developed Google Earth Engine and Python scripts as free and open-source software.

Hybrid model to optimize object-based land cover classification by meta-heuristic algorithm: an example for supporting urban management in Ha Noi, Viet Nam

ABSTRACTThis study proposed a novel object-based hybrid classification model named GMNN that combines Grasshopper Optimization Algorithm (GOA) and the multiple-class Neural network (MNN) for urban pattern detection in Hanoi, Vietnam. Four bands of SPOT 7 image and derivable NDVI, NDWI were used to generate image segments with associated attributes by PCI Geomatics software. These segments were classified into four urban surface types (namely water, impervious surface, vegetation and bare soil) by the proposed model. Alternatively, three training and validation datasets of different sizes were used to verify the robustness of this model. For all tests, the overall accuracies of the classification were approximately 87%, and the Area under Receiver Operating Characteristic curves for each land cover type was 0.97. Also, the performance of this model was examined by comparing several statistical indicators with common benchmark classifiers. The results showed that GMNN out-performed established methods in all comparable indicators. These results suggested that our hybrid model was successfully deployed in the study area and could be used as an alternative classification method for urban land cover studies. In a broader sense, classification methods will be enriched with the active and fast-growing contribution of metaheuristic algorithms.

A fully learnable context-driven object-based model for mapping land cover using multi-view data from unmanned aircraft systems

Context information is rarely used in the object-based landcover classification. Previous models that attempted to utilize this information usually required the user to input empirical values for critical model parameters, leading to less optimal performance. Multi-view image information is useful for improving classification accuracy, but the methods to assimilate multi-view information to make it usable for context driven models have not been explored in the literature. Here we propose a novel method to exploit the multi-view information for generating class membership probability. Moreover, we develop a new conditional random field model to integrate multi-view information and context information to further improve landcover classification accuracy. This model does not require the user to manually input parameters because all parameters in the Conditional Random Field (CRF) model are fully learned from the training dataset using the gradient descent approach. Using multi-view data extracted from small Unmanned Aerial Systems (UASs), we experimented with Gaussian Mixed Model (GMM), Random Forest (RF), Support Vector Machine (SVM) and Deep Convolutional Neural Networks (DCNN) classifiers to test model performance. The results showed that our model improved average overall accuracies from 58.3% to 74.7% for the GMM classifier, 75.8% to 87.3% for the RF classifier, 75.0% to 84.4% for the SVM classifier and 80.3% to 86.3% for the DCNN classifier. Although the degree of improvement may depend on the specific classifier respectively, the proposed model can significantly improve classification accuracy irrespective of classifier type.

Analysis of Land Cover Classification Using Multi-Wavelength LiDAR System

The airborne multi-wavelength light detection and ranging (LiDAR) system measures different wavelengths simultaneously and usually includes two or more active channels in infrared and green to acquire both topographic and hydrographic information. The reflected multi-wavelength energy can also be used to identify different land covers based on physical properties of materials. This study explored the benefits of multi-wavelength LiDAR in object-based land cover classification, focusing on three major issues: (1) the evaluation of single- and multi-wavelength LiDARs for land cover classification; (2) the performance of spectral and geometrical features extracted from multi-wavelength LiDAR; and (3) the comparison of the vegetation index derived from active multi-wavelength LiDAR and passive multispectral images. The three-wavelength test data were acquired by Optech Titan in green, near-infrared, and mid-infrared channels, and the reference data were acquired from Worldview-3 image. The experimental results show that the multi-wavelength LiDAR provided higher accuracy than single-wavelength LiDAR in land cover classification, with an overall accuracy improvement rate about 4–14 percentage points. The spectral features performed better compared to geometrical features for grass, road, and bare soil classes, and the overall accuracy improvement is about 29 percentage points. The results also demonstrated the vegetation indices from Worldview-3 and Optech Titan have similar characteristics, with correlations reaching 0.68 to 0.89. Overall, the multi-wavelength LiDAR system improves the accuracy of land cover classification because this system provides more spectral information than traditional single-wavelength LiDAR.

Object-based analysis of multispectral airborne laser scanner data for land cover classification and map updating

During the last 20years, airborne laser scanning (ALS), often combined with passive multispectral information from aerial images, has shown its high feasibility for automated mapping processes. The main benefits have been achieved in the mapping of elevated objects such as buildings and trees. Recently, the first multispectral airborne laser scanners have been launched, and active multispectral information is for the first time available for 3D ALS point clouds from a single sensor. This article discusses the potential of this new technology in map updating, especially in automated object-based land cover classification and change detection in a suburban area. For our study, Optech Titan multispectral ALS data over a suburban area in Finland were acquired. Results from an object-based random forests analysis suggest that the multispectral ALS data are very useful for land cover classification, considering both elevated classes and ground-level classes. The overall accuracy of the land cover classification results with six classes was 96% compared with validation points. The classes under study included building, tree, asphalt, gravel, rocky area and low vegetation. Compared to classification of single-channel data, the main improvements were achieved for ground-level classes. According to feature importance analyses, multispectral intensity features based on several channels were more useful than those based on one channel. Automatic change detection for buildings and roads was also demonstrated by utilising the new multispectral ALS data in combination with old map vectors. In change detection of buildings, an old digital surface model (DSM) based on single-channel ALS data was also used. Overall, our analyses suggest that the new data have high potential for further increasing the automation level in mapping. Unlike passive aerial imaging commonly used in mapping, the multispectral ALS technology is independent of external illumination conditions, and there are no shadows on intensity images produced from the data. These are significant advantages in developing automated classification and change detection procedures.

Quantitative assessment of 2014–2015 land-cover changes in Azerbaijan using object-based classification of LANDSAT-8 timeseries

The main goals of this study are the object-based land-cover classification of LANDSAT-8 satellite imagery of 2014 and 2015, the quantitative assessment of gross and net changes of agricultural land, built-up areas, forest, bare soil and forest between 2014 and 2015, the quantification of the Normalized Difference Vegetation Index (NDVI) rates within these land-cover classes, and the change detection analysis between the NDVIs. The achieved overall accuracies of object-based classification for the 2014 and the 2015 land-cover maps were 82 and 87 %, respectively. Therefore, the achieved accuracies were considered to be acceptable for quantified change detection analyses. For the gross areas of agricultural land, forest and built-up areas an increase was observed. The agricultural gross area was 30,911 km2 in 2014 and 31,999 km2 in 2015. The gross area of the built-up land increased from 12,550 to 13,548 km2. The gross area of forest land changed from 8211 to 9175 km2. A decrease was observed in the gross area of grassland from 28,229 to 24,925 km2. This was primarily related to the land-cover shifts driven by agricultural activities. The gross areas of bare soil and water bodies did not change significantly. The net change analysis, however, revealed significant differences in comparison to gross change areas for both gains and losses of the land-cover classes. The net change analysis revealed positive net changes of 7229, 5576, 1337, 399, 951 km2 for agricultural land, forest, built-up areas, bare soil and water bodies, correspondingly. A negative net change of 2198 km2 was observed for grassland. This allows to conclude that the negative net change of grassland was related with the significant changes of grassland into agricultural land. No significant net changes were observed for the bare soil land-cover class. The classification of NDVIs derived from 2014 to 2015 LANDSAT-8 OLI satellite images showed that the vegetation cover of agricultural and built-up land-cover increased for the low (0.1–0.2) and medium (0.2–0.3) and decreased for the high NDVI values (0.3–1). The area of high (0.3–1) NDVIs in the forest land-cover was observed to be higher in 2015 than in 2014. A reduction in the low (0.1–0.2), medium (0.2–0.3) and high NDVI values (0.3–1) was observed for the grasslands land-cover. The reductions of the high NDVI rates (0.3–1) observed for agricultural, build-up and grasslands land-cover types may be related to agricultural and industrial activities and also to climate change impacts. For the entire coverage of Azerbaijan, positive and negative NDVI changes of 3170 and 3859 km2 respectively were observed.

Integrating User Needs on Misclassification Error Sensitivity into Image Segmentation Quality Assessment

AbstractCommonly the assessment of the quality of image segmentations used in object-based land cover classification uses the geometric match between the derived segmentation and a reference dataset. This paper argues that a more appropriate assessment of a segmentation is to also consider the thematic content of the objects generated. This allows the assessment to be tailored to the needs of the specific user. A new method for image segmentation quality assessment is described, which combines a traditional geometric-only method with the thematic similarity index (TSI), a metric that expresses the degree of thematic quality of objects from a user’s perspective. The perspectives of two users (a wolf researcher and a general user of land cover information) were adopted in a case study to demonstrate the new method. The results show that the new method allowed the production of more accurate land cover classifications for the two users than the use of the geometric-only approach.

Using object-based hierarchical classification to extract land use land cover classes from high-resolution satellite imagery in a complex urban area

Producing land use and land cover (LULC) maps, particularly in complex urban areas, is one of the most important necessities in civil management programs and is an important research topic in satellite image analysis. High-resolution satellite images provide more opportunities for cost-benefit production of such information. This paper proposes a hierarchical LULC classification based on image objects that are created from multiresolution segmentation. A rule-based strategy is used to implement a step-by-step object-based land cover classification on a pan-sharpened IKONOS image taken from a complex urban region in Shiraz, Iran. A new spatial geometrical analysis for the reclassification of unclassified land cover objects is also utilized. After the initial classification, an object-based land use classification is implemented based on the land cover results and using conceptual, spatial, and geometrical modeling of the relationships between land use elements. Overall classification accuracy was 89 and 87% for land cover and land use approaches, respectively. In the best unclassified object analysis, ∼70% of unclassified objects were reclassified correctly. The hierarchical methodology proposed here results in fewer unclassified objects since a multistage classification process is utilized rather than the traditional one-pass classification.

Remote sensing and object-based techniques for mapping fine-scale industrial disturbances

Remote sensing provides an important data source for the detection and monitoring of disturbances; however, using this data to recognize fine-spatial resolution industrial disturbances dispersed across extensive areas presents unique challenges (e.g., accurate delineation and identification) and deserves further investigation. In this study, we present and assess a geographic object-based image analysis (GEOBIA) approach with high-spatial resolution imagery (SPOT 5) to map industrial disturbances using the oil sands region of Alberta's northeastern boreal forest as a case study. Key components of this study were (i) the development of additional spectral, texture, and geometrical descriptors for characterizing image-objects (groups of alike pixels) and their contextual properties, and (ii) the introduction of decision trees with boosting to perform the object-based land cover classification. Results indicate that the approach achieved an overall accuracy of 88%, and that all descriptor groups provided relevant information for the classification. Despite challenges remaining (e.g., distinguishing between spectrally similar classes, or placing discrete boundaries), the approach was able to effectively delineate and classify fine-spatial resolution industrial disturbances.

Object-based land-cover classification for metropolitan Phoenix, Arizona, using aerial photography

Detailed land-cover mapping is essential for a range of research issues addressed by the sustainability and land system sciences and planning. This study uses an object-based approach to create a 1m land-cover classification map of the expansive Phoenix metropolitan area through the use of high spatial resolution aerial photography from National Agricultural Imagery Program. It employs an expert knowledge decision rule set and incorporates the cadastral GIS vector layer as auxiliary data. The classification rule was established on a hierarchical image object network, and the properties of parcels in the vector layer were used to establish land cover types. Image segmentations were initially utilized to separate the aerial photos into parcel sized objects, and were further used for detailed land type identification within the parcels. Characteristics of image objects from contextual and geometrical aspects were used in the decision rule set to reduce the spectral limitation of the four-band aerial photography. Classification results include 12 land-cover classes and subclasses that may be assessed from the sub-parcel to the landscape scales, facilitating examination of scale dynamics. The proposed object-based classification method provides robust results, uses minimal and readily available ancillary data, and reduces computational time.

Impact of Training Set Size on Object-Based Land Cover Classification

Supervised classifiers are commonly employed in remote sensing to extract land cover information, but various factors affect their accuracy. The number of available training samples, in particular, is known to have a significant impact on classification accuracies. Obtaining a sufficient number of samples is, however, not always practical. The support vector machine (SVM) is a supervised classifier known to perform well with limited training samples and has been compared favourably to other classifiers for various problems in pixel-based land cover classification. Very little research on training-sample size and classifier performance has been done in a geographical object-based image analysis (GEOBIA) environment. This paper compares the performance of SVM, nearest neighbour (NN) and maximum likelihood (ML) classifiers in a GEOBIA environment, with a focus on the influence of training-set size. Training-set sizes ranging from 4-20 per land cover class were tested. Classification tree analysis (CTA) was used for feature selection. The results indicate that the performance of all the classifiers improved significantly as the size of the training set increased. The ML classifier performed poorly when few (&lt;10 per class) training samples were used and the NN classifier performed poorly compared to SVM throughout the experiment. SVM was the superior classifier for all training-set sizes although ML achieved competitive results for sets of 12 or more training areas per class.

Spectral normalization of SPOT 4 data to adjust for changing leaf phenology within seasonal forests in Cambodia

As cloud cover exacerbates the application of optical satellite data for forest monitoring in tropical wet and dry regions during the rainy season, data acquisition is mainly restricted to the dry season. When analyzing wide areas, large numbers of single scenes obtained at different times of the dry season are often handled. Such imagery is characterized by changes of spectral reflectance due to vegetation phenology, varying atmospheric effects and solar geometries. In order to allow batch processing with automatic classification techniques, inter-scene comparability is required and data have to be radiometrically normalized. Cambodia is characterized by a mixture of evergreen, semi-evergreen and deciduous forest types, the latter two experiencing at least partial leaf shedding over the course of the dry season. Using spatial medium resolution SPOT 4 data and a manually delineated base map a season adjustment model was developed. The model is adapting the land cover specific spectral signatures of a slave scene (acquired in the middle of the dry season with its seasonal forests defoliated) to an adjacent master scene (from the beginning of the dry season, showing the same forest types with leafs). The relative position of every pixel reflectance was determined in relation to the mean reflectance and its standard deviation for each land cover type and sensor band of the unadjusted slave scene. For seasonality adjustment these pixel reflectance values were transformed (rescaled) to the corresponding position in spectral space defined by the band mean reflectance and standard deviation derived from the corresponding land cover class of the master scene. While the variability of spectral profiles of the pixels in the slave scene is rescaled, the mean reflectance value of the land cover class in the slave scene is conformed to the mean reflectance of the corresponding land cover class in the master scene. The Transformed Divergence (TD) separability index was used to indicate the performance of the adjustment process by characterizing the spectral distance for each land cover type comparing a reference dataset to the uncorrected and to the seasonality corrected scene respectively. While the TD values of all forest types showed a sharp decline, highlighting the good performance of the model, the TD values of the agriculture/urban class remained high, indicating limited normalization of this heterogeneous land cover type. In order to further demonstrate the performance of the model, an object-based land cover classification was applied to the unadjusted as well as to the corresponding adjusted scene. A comparison of the results showed a highly significant improvement of overall accuracy from 32.2% to 75.8% when applying seasonality adjustment.