Object-oriented Classification Approach Research Articles

Fine Classification of UAV Urban Nighttime Light Images Based on Object-Oriented Approach.

Fine classification of urban nighttime lighting is a key prerequisite step for small-scale nighttime urban research. In order to fill the gap of high-resolution urban nighttime light image classification and recognition research, this paper is based on a small rotary-wing UAV platform, taking the nighttime static monocular tilted light images of communities near Meixi Lake in Changsha City as research data. Using an object-oriented classification method to fully extract the spectral, textural and geometric features of urban nighttime lights, we build four types of classification models based on random forest (RF), support vector machine (SVM), K-nearest neighbor (KNN) and decision tree (DT), respectively, to finely extract five types of nighttime lights: window light, neon light, road reflective light, building reflective light and background. The main conclusions are as follows: (i) The equal division of the image into three regions according to the visual direction can alleviate the variable scale problem of monocular tilted images, and the multiresolution segmentation results combined with Canny edge detection are more suitable for urban nighttime lighting images; (ii) RF has the highest classification accuracy among the four classification algorithms, with an overall classification accuracy of 95.36% and a kappa coefficient of 0.9381 in the far view region, followed by SVM, KNN and DT as the worst; (iii) Among the fine classification results of urban light types, window light and background have the highest classification accuracy, with both UA and PA above 93% in the RF classification model, while road reflective light has the lowest accuracy; (iv) Among the selected classification features, the spectral features have the highest contribution rates, which are above 59% in all three regions, followed by the textural features and the geometric features with the smallest contribution rates. This paper demonstrates the feasibility of nighttime UAV static monocular tilt image data for fine classification of urban light types based on an object-oriented classification approach, provides data and technical support for small-scale urban nighttime research such as community building identification and nighttime human activity perception.

INTRODUCTION DE LANACARDIER EN ZONE SOUDANIENNE ET EVALUATION DE SON IMPACT SUR LA DYNAMIQUE DE LOCCUPATION DES SOLS DE 1986 A 2022 (COTE DIVOIRE)

The objective of this project is to apply to some optical data, an object oriented classification approach for monitoring and detecting the changes of land cover in the north of Cote dIvoire where cashew is spreading. Two Landsat images TM (1986)/ETM+ (2002) and a Sentinel-2 (2022) are used in this research. The method is achieved through the monitoring of land cover from an object-oriented approach based on fuzzy logic. For this specific methodological approach, we resort to texture analysis and principal component analysisa change detection through a post classification method and the integration of the resulting map in a GIS. By this approach, we produced the lands cover of 1986 2002 and 2022 with Kappa coefficients at 86 88 and 87 % respectively. For the period of concern (1986-2022), the method allows spatializing and quantifying the land cover changes in this complex environment dominated by cashew.

Burn characterization using object-oriented hyperspectral image classification.

This paper presents a new approach based on hyperspectral imaging combined with an object-oriented classification method that allows the generation of burn depth classification maps facilitating easier characterization of burns. Hyperspectral images of 14 patients diagnosed with burns on the upper and lower limbs were acquired using a pushbroom hyperspectral imaging system. The images were analyzed using an object-oriented classification approach that uses objects with specific spectral, textural and spatial attributes as the minimum unit for classifying information. The method performance was evaluated in terms of overall accuracy, sensitivity, precision and specificity computed from the confusion matrix. The results revealed that the approach proposed in this study performed well in differentiating burn classes with a high level of overall accuracy (95.99% ± 0.60%), precision (97.30% ± 2.46%), sensitivity (97.23% ± 3.02%) and specificity (98.02% ± 1.98%). In conclusion, the object-based approach for burns hyperspectral images classification can provide maps that can help surgeons identify with better precision different depths of burn wounds.

Detailed Mapping of Lava and Ash Deposits at Indonesian Volcanoes by Means of VHR PlanetScope Change Detection

Mapping of lava flows in unvegetated areas of active volcanoes using optical satellite data is challenging due to spectral similarities of volcanic deposits and the surrounding background. Using very high-resolution PlanetScope data, this study introduces a novel object-oriented classification approach for mapping lava flows in both vegetated and unvegetated areas during several eruptive phases of three Indonesian volcanoes (Karangetang 2018/2019, Agung 2017, Krakatau 2018/2019). For this, change detection analysis based on PlanetScope imagery for mapping loss of vegetation due to volcanic activity (e.g., lava flows) is combined with the analysis of changes in texture and brightness, with hydrological runoff modelling and with analysis of thermal anomalies derived from Sentinel-2 or Landsat-8. Qualitative comparison of the mapped lava flows showed good agreement with multispectral false color time series (Sentinel-2 and Landsat-8). Reports of the Global Volcanism Program support the findings, indicating the developed lava mapping approach produces valuable results for monitoring volcanic hazards. Despite the lack of bands in infrared wavelengths, PlanetScope proves beneficial for the assessment of risk and near-real-time monitoring of active volcanoes due to its high spatial (3 m) and temporal resolution (mapping of all subaerial volcanoes on a daily basis).

Crop Phenology-Based, Object-Oriented Classification Approach Using SENTINEL-2A and NDVI Time Series: Sunflower Crops in Kırklareli TURKEY

The aim of this study is to develop a methodology for determining sunflower cultivated areas with the help of high resolution SENTINEL-2A satellite images time series representing the phenological stages of the crop growth cycle, and its application in Kırklareli province. Spectral information representing phenological periods was obtained with the help of satellite images and normalized difference vegetation index (NDVI) time series, and an object-oriented classification approach was developed based on this spectral information database. Segmentation and classification decision tree algorithms were produced by using this spectral information database, object shape criteria and other auxiliary thematic maps. The best performance in segmentation was achieved by increasing the weight coefficient of the "Canny edge” layer, which is the edge determination layer defined in the multiresolution method of "Canny edge” algorithm to define the agricultural parcels. Object-oriented classification was carried out based on the this segmented parcels. First, summer, winter, fallow and continuous green areas were determined through the classification decision tree algorithms. The summer and winter crops were classified using the parcel spectral information of the crop-based learning samples that allocated in field work. The crops for which class definition could not be made were passed through a second elimination in the "unclassified" group and later assigned to their classes. In the last stage, parcels whose class definition could not be made were named as "other" class. According to the confusion matrix and accuracy analysis results, sunflower, which was determined in two classes as early and late sowing, was classified at 98% and 92% accuracy, respectively.

Mapping flood by the object-based method using backscattering coefficient and interference coherence of Sentinel-1 time series

The SAR has the ability of all-weather and all-time data acquisition, it can penetrate the cloud and is not affected by extreme weather conditions, and the acquired images have better contrast and rich texture information. This paper aims to investigate the use of an object-oriented classification approach for flood information monitoring in floodplains using backscattering coefficients and interferometric coherence of Sentinel-1 data under time series. Firstly, the backscattering characteristics and interference coherence variation characteristics of SAR time series are used to analyze whether the flood disaster information can be accurately reflected and provide the basis for selecting input classification characteristics of subsequent SAR images. Subsequently, the contribution rate index of the RF model is used to calculate the importance of each index in time series to convert the selected large number of classification features into low dimensional feature space to improve the classification accuracy and reduce the data redundancy. Finally, the SAR image features in each period after multi-scale segmentation and feature selection are jointly used as the input features of RF classification to extract and segment the water in the study area to monitor floods' spatial distribution and dynamic characteristics. The results showed that the various attributes of backscatter coefficients and interferometric coherence under time series could accurately correspond with the actual flood risk, and the combined use of backscattering coefficient and interferometric coherence for flood extraction can significantly improve the accuracy of flood information extraction. Overall, the object-based random forest method using the backscattering coefficient and interference coherence of Sentinel-1 time series for flood extraction advances our understanding of flooding's temporal and spatial dynamics, essential for the timely adoption of adaptation and mitigation strategies for loss reduction.

Formalized control logic fault definition with ontological reasoning for air handling units

Control logic programs determine the behavior of Heating, Ventilation, and Air Conditioning (HVAC) systems under different operating conditions. Faults in control logic account for more than 15% of all HVAC system problems, causing energy waste and occupancy discomfort. The first step towards systematically detecting and diagnosing control logic faults is to have an unambiguous control logic fault definition so that the existence of control logic faults in an HVAC system can be discovered from its operational data. In order to have high level of fault detection accuracy, the control logic fault definition needs to be customized for different HVAC systems as they have different components and control sequences. In this paper, we propose an object-oriented classification approach to systematically define customized control logic faults in terms of control logic input/output variable expressions. Focusing on air handling unit (AHU) systems, and their general operation objectives of energy efficiency and occupancy comfort, we elaborated four control goals and developed corresponding reasoning mechanisms to derive fault definitions. We also developed an HVAC component and control information ontology to be used in these reasoning mechanisms by extending existing HVAC information models. The prototype of the developed approach was tested with 27 common AHUs specified by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), and the results show that using the developed approach, it is possible to define a customized set of control logic faults applicable to each specific AHU with an overall precision of 94.2% and recall of 83.0%. This demonstrates the generality of our proposed approach in providing customized control logic fault definitions for different types of AHUs.

Coastal Wetland Classification with GF-3 Polarimetric SAR Imagery by Using Object-Oriented Random Forest Algorithm.

When the use of optical images is not practical due to cloud cover, Synthetic Aperture Radar (SAR) imagery is a preferred alternative for monitoring coastal wetlands because it is unaffected by weather conditions. Polarimetric SAR (PolSAR) enables the detection of different backscattering mechanisms and thus has potential applications in land cover classification. Gaofen-3 (GF-3) is the first Chinese civilian satellite with multi-polarized C-band SAR imaging capability. Coastal wetland classification with GF-3 polarimetric SAR imagery has attracted increased attention in recent years, but it remains challenging. The aim of this study was to classify land cover in coastal wetlands using an object-oriented random forest algorithm on the basis of GF-3 polarimetric SAR imagery. First, a set of 16 commonly used SAR features was extracted. Second, the importance of each SAR feature was calculated, and the optimal polarimetric features were selected for wetland classification by combining random forest (RF) with sequential backward selection (SBS). Finally, the proposed algorithm was utilized to classify different land cover types in the Yancheng Coastal Wetlands. The results show that the most important parameters for wetland classification in this study were Shannon entropy, Span and orientation randomness, combined with features derived from Yamaguchi decomposition, namely, volume scattering, double scattering, surface scattering and helix scattering. When the object-oriented RF classification approach was used with the optimal feature combination, different land cover types in the study area were classified, with an overall accuracy of up to 92%.

Object-oriented classification approach for bone metastasis mapping from whole-body bone scintigraphy.

Whole-body bone scintigraphy is the most widely used method for detecting bone metastases in advanced cancer. However, its interpretation depends on the experience of the radiologist. Some automatic interpretation systems have been developed in order to improve diagnostic accuracy. These systems are pixel-based and do not use spatial or textural information of groups of pixels, which could be very important for classifying images with better accuracy. This paper presents a fast method of object-oriented classification that facilitates easier interpretation of bone scintigraphy images. Nine whole-body images from patients suspected with bone metastases were analyzed in this preliminary study. First, an edge-based segmentation algorithm together with the full lambda-schedule algorithm were used to identify the object in the bone scintigraphy and the textural and spatial attributes of these objects were calculated. Then, a set of objects (224 objects, ~ 46% of the total objects) were selected as training data based on visual examination of the image, and were assigned to various levels of radionuclide accumulation before performing the data classification using both k-nearest-neighbor and support vector machine classifiers. The performance of the proposed method was evaluated using as metric the statistical parameters calculated from error matrix. The results revealed that the proposed object-oriented classification approach using either k-nearest-neighbor or support vector machine as classification methods performed well in detecting bone metastasis in terms of overall accuracy (86.62±2.163% and 86.81±2.137% respectively) and kappa coefficient (0.6395±0.0143 and 0.6481±0.0218 respectively). In conclusion, the described method provided encouraging results in mapping bone metastases in whole-body bone scintigraphy.

Road Extraction from Cartosat-2F Multispectral Data with Object-Oriented Analysis

For detection of a road network, high-resolution satellite data have been used following the object-oriented classification approach. We used object-based feature extraction algorithms for detection of road networks from a high resolution Cartosat-2F multispectral data in an Indian city with varying terrain conditions ranging from a compact built-up area to a predominantly vegetated area. The approach involves multi-resolution segmentation (MRS ) and spectral difference segmentation (SDS ) followed by road extraction using fuzzy rule-based algorithm based on various object features, viz. gray-level co-occurrence matrix homogeneity, density, rectangular fit, etc. With overall accuracies ranging from 77.46% to 92% SDS approach performed better than MRS which could afford 60.46% to 75.0% only. However, both of these approaches score over the classical Gaussian maximum likelihood classifier which could register only 50.0% to 68.0% overall accuracy. Furthermore, the maximum overall accuracy was obtained in compact built-up site (85% to 92%) followed by sparsely built-up site (75.0% to 88%) and predominantly vegetated site (60.46% to 77.14%).

Landslide Mapping Using Multiscale LiDAR Digital Elevation Models

ABSTRACT This study presents a new methodology to identify landslide and landslide-susceptible locations in Interior Alaska using only geomorphic properties from light detection and ranging (LiDAR) derivatives (i.e., slope, profile curvature, and roughness) and the normalized difference vegetation index (NDVI), focusing on the effect of different resolutions of LiDAR images. We developed a semi-automated object-oriented image classification approach in ArcGIS 10.5 and prepared a landslide inventory from visual observation of hillshade images. The multistage work flow included combining derivatives from 1-, 2.5-, and 5-m-resolution LiDAR, image segmentation, image classification using a support vector machine classifier, and image generalization to clean false positives. We assessed classification accuracy by generating confusion matrix tables. Analysis of the results indicated that LiDAR image scale played an important role in the classification, and the use of NDVI generated better results. Overall, the LiDAR 5-m-resolution image with NDVI generated the best results with a kappa value of 0.55 and an overall accuracy of 83 percent. The LiDAR 1-m-resolution image with NDVI generated the highest producer accuracy of 73 percent in identifying landslide locations. We produced a combined overlay map by summing the individual classified maps that was able to delineate landslide objects better than the individual maps. The combined classified map from 1-, 2.5-, and 5-m-resolution LiDAR with NDVI generated producer accuracies of 60, 80, and 86 percent and user accuracies of 39, 51, and 98 percent for landslide, landslide-susceptible, and stable locations, respectively, with an overall accuracy of 84 percent and a kappa value of 0.58. This semi-automated object-oriented image classification approach demonstrated potential as a viable tool with further refinement and/or in combination with additional data sources.

THE IMPACT OF FOREST FIRE ON FOREST COVER TYPES IN MONGOLIA

Abstract. The objective of this study was the impact of forest fire on forest cover types. This study has identified non-forest and forest area that has seven forest class are included with cedar, pine, larch, birch, birch-pine mixed, birch-larch mixed and cedar-larch mixed, additionally, remote sensing imagery is applied. In contrast, Landsat imagery has been used several classification approaches. Moreover, the current classification has developments in segmentation and object-oriented techniques offer the suitable analysis to classify satellite data. In the object-oriented classification approach, images cluster to homogenous area as forest types by suitable parameters in some level. The accuracy analysis revealed that overall accuracy showed a good accuracy of determination (86.33 percent in 2000 and 93.75 percent in 2011) with regard to identify of the forest cover and type. Furthermore, these results suggest that the Landsat TM and ETM+ data can reliable detect the forest type based upon the segmentation and object-oriented techniques. In generally, our study area is high-risky region to forest fires. It is higher influence to forest cover and tree species and other ecosystems. Overall, wildfire of impact results showed that 25239 ha of forests were changed to burnt area and 52603 ha forests were changed to grassland.

CNN BASED DETECTION OF BUILDING ROOFS FROM HIGH RESOLUTION SATELLITE IMAGES

Abstract. The detection and reconstruction of building have attracted more attention in the community of remote sensing and computer vision. Light detection and ranging (LiDAR) has been proved to be a good way to extract building roofs, while we have to face the problem of data shortage for most of the time. In this paper, we tried to extract the building roofs from very high resolution (VHR) images of Chinese satellite Gaofen-2 by employing convolutional neural network (CNN). It has been proved that the CNN is of a higher capability of recognizing detailed features which may not be classified out by object-based classification approach. Several major steps are concerned in this study, such as generation of training dataset, model training, image segmentation and building roofs recognition. First, urban objects such as trees, roads, squares and buildings were classified based on random forest algorithm by an object-oriented classification approach, the building regions were separated from other classes at the aid of visually interpretation and correction; Next, different types of building roofs mainly categorized by color and size information were trained using the trained CNN. Finally, the industrial and residential building roofs have been recognized individually and the results have been validated individually. The assessment results prove effectiveness of the proposed method with approximately 91% and 88% of quality rates in detection industrial and residential building roofs, respectively. Which means that the CNN approach is prospecting in detecting buildings with a very higher accuracy.

NATIONAL LEVEL INVENTORY OF COFFEE PLANTATIONS USING HIGH RESOLUTION SATELLITE DATA

Abstract. Coffee is the second most traded commodity in the world and its production has implications in both international and domestic economy. It is an important commercial crop of India and hence, reliable acreage and production estimation is most essential for taking up policy decisions. The coffee growing regions in India are mainly confined to the traditional South Indian states (Karnataka, Kerala and Tamil Nadu) and partly in non-traditional regions (Andhra Pradesh and Odisha) while to a smaller extent in North-Eastern states. Interpretation and mapping of coffee plantations using satellite data is quite challenging due to the diverse and complex cultivation practices. In the present study, multi-resolution and multi-source data was utilized for mapping of coffee plantations in the country. Temporal LISS-III (24.0 m) data was used for characterizing the phenology of coffee and other competing plantation crops for selection of optimal high resolution satellite (HRS) datasets. Accordingly, Cartosat-1 (2.5 m) and Resourcesat LISS-IV multispectral (5.0 m) datasets corresponding to February-April months were utilized. The spectral signature of coffee plantations is determined by the age category of coffee plantations, varietal difference, density & composition of shade trees along with terrain features like slope and aspect. The plantations manifested in different tones of red and mottled texture on the multispectral image. Object oriented classification approach showed encouraging results in homogenous & contiguous areas but showed poor mapping accuracy in heterogeneous regions due to complex spectral signature and varying texture. Thus, a combination of digital and visual interpretation techniques were used for mapping of coffee plantations depending on the suitability. Feature space optimization function was used for selection of object parameters and 14 image features consisting of mean spectral values, standard deviation, NDVI, geometry and contextual parameters were used for classification of coffee plantations using Support Vector Machine (SVM). In case of small holdings and heterogeneous areas, interactive visual interpretation of HRS data at 1 : 5,000 scale using tone, texture, shape and terrain characteristics was carried out for mapping of coffee plantations with the help of ground truth and field experience of Liaison Officials of Coffee Board. Post-interpretation field verification/validation of the interpreted maps was carried out for the accuracy assessment and the overall mapping accuracy of better than 90.0 per cent was achieved in the study. Total area under coffee plantations was about 4.41 lakh ha (excluding N-E states). This is the first study in the country to generate coffee map at national level for creation of baseline geospatial database which could be updated periodically. Suitability analysis using pedo-climatic and terrain parameters is being carried to promote the coffee cultivation in the non-traditional regions. Further research efforts are necessary for varietal discrimination and modelling of pests and diseases which is critical input for production estimation.

Monitoring and predicting land use/cover changes in the Aksu-Tarim River Basin, Xinjiang-China (1990–2030)

Land use/cover (LCLU) is considered as one of the most serious environmental challenges that threatens developed and less developed countries. LCLU changes' monitoring using the integration of remote sensing (RS) and geographical information systems (GIS) and their predicting using an artificial neural network (ANN) in the western part of the Tarim River Basin (Aksu), north-western Xinjiang-China, from 1990 to 2030 have been investigated first time through satellite imageries available. The imageries of 1990, 2000, 2005, 2010, and 2015 were downloaded from GLCF and USGS websites. After digital image processing, the object-oriented image classification approach was applied. The ANN method with MOLUSCE Plugin was used to simulate the LCLU changes in 2020, 2025, and 2030. GIS has also been used to calculate the distance from the road and water and etc. The simulation results of 2010 and 2015 were validated using classification data with Kappa coefficient. The results showed high accuracy of the classification and prediction as the validation of simulated 2010 and 2015 maps to the referenced maps have high accuracy of Kappa 84 and 88%, respectively. The results revealed that the land cover classes forest-, grass-, wet-, and barren land have been decreased from 50.01, 13.06, 8.24, and 1.06% in 1990 to 32.03, 3.06, 6.26, and 0.97% in 2015, respectively, while the land use classes, crop or farm land, and urban land have been increased almost double from 25.5 and 2.13% in 1990 to 53.71 and 3.86% from the total area in 2015, respectively. For the prediction, forest- and wetlands will loss more than half of their areas by 2030, the grass land will be cleared completely to be only 1.3% from the total study area, while the urban land will be increased to be 4.4% or the double of 1990. These results are attributed to population growth and expanding of agriculture land on the grass land, but the effect of climate was weak as the rainfall increased during the study period. Causes and effects of the LCLU changes were briefly discussed. The output of the study serves as useful tools for policy and decision makers combatting natural resources misused in arid lands.

Multitemporal variables for the mapping of coffee cultivation areas

Abstract: The objective of this work was to propose a new methodology for mapping coffee cropping areas that includes multitemporal data as input parameters in the classification process, by using the Landsat TM NDVI time series, together with an object-oriented classification approach. The algorithm BFAST was used to analyze coffee, pasture, and native vegetation temporal profiles, allied to a geographic object-based image analysis (GEOBIA) for mapping. The following multitemporal variables derived from the R package greenbrown were used for classification: mean, trend, and seasonality. The results showed that coffee, pasture, and native vegetation have different temporal behaviors, which corroborates the use of these data as input variables for mapping. The classifications using temporal variables, associated with spectral data, achieved high-global accuracy rates with 93% hit. When using only temporal data, ratings also showed a hit percentage above 80% accuracy. Data derived from Landsat TM time series are efficient for mapping coffee cropping areas, reducing confusion between targets and making the classification process more accurate, contributing to a correct characterization and mapping of objects derived from a RapidEye image, with a high spatial solution.

Spatial-Temporal Analysis and Stability Investigation of Coastline Changes: A Case Study in Shenzhen, China

Coastlines are some of the most important dynamic linear features on the Earth's surface. Conducting research on coastline changes is essential to improve the environmental management and development of the coast. In this study, Landsat satellite images that cover the entire coastal zone of Shenzhen, which was the first special economic zone in China, are used to study coastline changes. Multitemporal coastline information for Shenzhen from 1988 to 2015 is extracted based on the object-oriented classification approach. The overall classification accuracy of this approach can reach 85% or more when it is used to classify the land-use types present within the coastal zone in this study. Moreover, coastline change rates are calculated using a point-based approach. Coastline stability changes are analyzed, and driving rules are explored, using a data-mining approach that addresses natural and social factors. The research results demonstrate that the object-based approach can extract accurate coastline information from Landsat images and that the coastlines in Shenzhen have varied with time and location. The coastline changes in Shenzhen from 1988 to 2015 can be divided into two stages: one displays accelerated change (1988–2003), whereas the other displays decelerated change (2003–2015). Completely different change characteristics of coastline stability are found between the eastern and western coastlines in Shenzhen, and the differences are mainly reflected in the coastline's morphological changes and varying rules.

Land Use and Land Cover Change in Vientiane Area, Lao PDR Using Object-Oriented Classification on Multi-Temporal Landsat Data

Monitoring of land use and land cover change using remote sensing is important to evaluate the impacts of anthropogenic activities on the environment. Digital change detection using post-classification can help to elucidate dynamics of landscape change. This study illustrates the effectiveness of object-oriented classification compared to pixel-oriented classification in generating land cover information and its temporal changes. Spatio-temporal dynamics of land cover types in Vientiane area, Lao PDR were analyzed using Landsat images in two-time series (1990 and 2015). We used the top-down approach to classify the Landsat images in iterative steps with three hierarchical scale levels. Scale levels of 25, 10 and 5 with different weighting parameters were used to map the land cover type of Vientiane in 1990 and 2015. With object-oriented classification, overall accuracy and Kappa statistic were improved by 13.44% and 0.16 for land cover classification (LCC) 1990. For LCC 2015, the improvements in overall accuracy and Kappa statistic were 28.71% and 0.25. Based on the LCC 1990 and 2015, we observed an significant growth of plantation areas over the 25 years in the study area . Instead of traditional agricultural activity, the plantation seemed to be the new driver in the rural areas of Lao PDR. The object-oriented classification approach can be applied in other areas of Lao PDR to generate accurate information on land cover changes for better land resource management.

Mapping urban impervious surfaces from an airborne hyperspectral imagery using the object-oriented classification approach

The objective of this research is to explore the capabilities of the hyperspectral imagery in mapping the urban impervious objects and identifying the surface materials using an object-oriented approach. The application is conducted to Toulouse city (France) within the HYEP research project in charge of using hyperspectral imagery for the environmental urban planning. The method uses the multi-resolution segmentation and classification algorithms. The first results highlight a high potential of the hyperspectral imagery in land cover mapping of the urban environment, especially the extraction of impervious surfaces. They, also, illustrate, that the object-oriented approach by means of the fuzzy logic classifier yields promising results in distinguishing the mean roofing materials based only on the spectral information. Conversely to the red clay tiles and metal roofs, which are easily identified, the concrete, gravel and asphalt roofs are still confused with roads.

Crop Classification Based on Feature Band Set Construction and Object-Oriented Approach Using Hyperspectral Images

Remote sensing plays a significant role for crop classification. Accurate crop classification is a common requirement to precision agriculture, including crop area estimation, crop yield estimation, precision crop management, etc. This paper developed a new crop classification method involving the construction and optimization of the vegetation feature band set (FBS) and combination of FBS and object-oriented classification (OOC) approach. In addition to the spectral and textural features of the original image, 20 spectral indices sensitive to the vegetation's biological parameters are added to the FBS to distinguish specific vegetation. A spectral dimension optimization algorithm of FBS based on class-pair separability (CPS) is also proposed to improve the separability between class pairs while reducing data redundancy. OOC approach is conducted on the optimized FBS based on CPS to reduce the salt-and-pepper noise. The proposed classification method was validated by two airborne hyperspectral images. The first image acquired in an agricultural area of Japan includes seven crop types, and the second image acquired in a rice breeding area consists of six varieties of rice. For the first image, the proposed method distinguished different vegetation with an overall accuracy of 97.84% and kappa coefficient of 0.96. For the second image, the proposed method distinguished the rice varieties accurately, achieving the highest overall accuracy (98.65%) and kappa coefficient (0.98). Results demonstrate that the proposed method can significantly improve crop classification accuracy and reduce edge effects, and that textural features combined with spectral indices sensitive to the chlorophyll, carotenoid, and Anthocyanin indicators contribute significantly to crop classification. Therefore, it is an effective approach for classifying crop species, monitoring invasive species, as well as precision agriculture related applications.