A spatialized classification approach for land cover mapping using hyperspatial imagery

An object-oriented, multiscale feature extraction approach is proposed for the land-cover classification of high spatial resolution images. The approach provides more discriminative features by considering the spatial context information from different segmentation levels. It consists of three successive substeps: segmentation by mean-shift algorithm, an iteratively merging process controlled by merging cost function and range-of-scale parameter, and feature extraction from linked multilevel image partitions. The mean-shift method is to get boundary-preserved and spectrally homogeneous over-segmentation regions. Then, a family of nested image partitions is constructed by a merging procedure. Meanwhile, every region of the finest scale is linked to image objects of its superlevels. Finally, every region in the finest scale is treated as a basic analysis unit, and the feature vectors are created by stacking statistics from the region and their superlevels. A support vector machine is used as a classifier and the method on two widely used high spatial resolution data sets over Pavia City, Italy, are evaluated. Compared with results reported in many papers, the result indicates superior accuracy.

The use of OBIA for high spatial resolution image classification can be divided in two main steps, the first being segmentation and the second regarding the labeling of the objects in accordance with a particular set of features and a classifier. Decision trees are often used to represent human knowledge in the latter. The issue falls in how to select a smaller amount of features from a feature space with spatial, spectral and textural variables to describe the classes of interest, which engenders the matter of choosing the best or more convenient feature selection (FS) method. In this work, an approach for FS within a decision tree was introduced using a single perceptron and the Backpropagation algorithm. Three alternatives were compared: single, double and multiple inputs, using a sequential backward search (SBS). Test regions were used to evaluate the efficiency of the proposed methods. Results showed that it is possible to use a single perceptron in each node, with an overall accuracy (OA) between 77.6% and 77.9%. Only SBS reached an OA larger than 88%. Thus, the quality of the proposed solution depends on the number of input features.

Spectral Properties and Heterogeneity of Phobos from Measurements byPhobos 2

Question: Can we reliably estimate grazing intensity, indicators of grazing intensity (i.e. field‐layer height and shrub‐cover), and vascular plant species richness in semi‐natural grasslands from high spatial resolution satellite data?Location: The Baltic Island of Öland (Sweden).Methods: Fieldwork included the on‐site description of grazed and ungrazed areas and shrub‐cover within 107 semi‐natural grassland sites. Field‐layer height and vascular plant species richness (total within‐site and mean small‐scale species richness) were recorded within the sites. Digital classification of QuickBird data was performed to identify grazed and ungrazed areas and shrub‐cover. Vegetation indices were generated to analyze the performance of satellite data for estimating field‐layer height, and the spectral heterogeneity was used to characterize the within‐site environmental heterogeneity.Results: The proportion of digitally classified grazed area explained 45% of the variation in field‐layer height and 43% of the variation in shrub‐cover. Field‐layer height was significantly related to vegetation indices. A linear model with three explanatory variables (spectral richnessred, spectral richnessNIR, and shrub‐cover) explained 47% of the variation in total within‐site species richness.Conclusions: High spatial resolution imagery may assist in the monitoring of the processes that follow the cessation of grazing, on the scale of individual grassland sites. Measures of spectral heterogeneity acquired by high spatial resolution imagery can be used in the assessment of total within‐site vascular plant species richness in semi‐natural grassland vegetation.

Remote sensing images have long been recognized as useful for the detection of building damages, mainly due to their wide coverage, revisit capabilities and high spatial resolution. The majority of contributions aimed at identifying debris and rubble piles, as the main focus is to assess collapsed and partially collapsed structures. However, these approaches might not be optimal for the image classification of façade damages, where damages might appear in the form of spalling, cracks and collapse of small segments of the façade. A few studies focused their damage detection on the façades using only post-event images. Nonetheless, several studies achieved better performances in damage detection approaches when considering multi-temporal image data. Hence, in this work a multi-temporal façade damage detection is tested. The first objective is to optimally merge pre- and post-event aerial oblique imagery within a supervised classification approach using convolutional neural networks to detect façade damages. The second objective is related to the fact that façades are normally depicted in several views in aerial manned photogrammetric surveys; hence, different procedures combining these multi-view image data are also proposed and embedded in the image classification approach. Six multi-temporal approaches are compared against 3 mono-temporal ones. The results indicate the superiority of multi-temporal approaches (up to ~25% in f1-score) when compared to the mono-temporal ones. The best performing multi-temporal approach takes as input sextuples (3 views per epoch, per façade) within a late fusion approach to perform the image classification of façade damages. However, the detection of small damages, such as smaller cracks or smaller areas of spalling, remains challenging in this approach, mainly due to the low resolution (~0.14 m ground sampling distance) of the dataset used.

Land cover change is a key factor influencing land surface temperature (LST) dynamics, particularly in coastal urban areas such as Pariaman City. This study aims to: (1) analyze land cover changes in Pariaman City in 2016 and 2023; (2) examine the spatial distribution of LST in both periods; and (3) evaluate the impact of land cover change on LST. A quantitative descriptive method was employed, using a supervised classification approach for land cover mapping, the Mono-Window Brightness Temperature model for LST estimation, and Pearson’s correlation and regression analyses to assess their relationship. The results indicate substantial land cover changes, with a forest area reduction of 26.162 km² and an increase in residential and agricultural land by 16.444 km² and 12.373 km², respectively. In 2016, LST distribution was dominated by the moderate class (20–24 °C), whereas in 2023 it exhibited greater variability, with an expansion of both low-temperature classes (<17 °C and 17–20 °C) and high-temperature classes (24–29 °C). Pearson’s correlation yielded r = −0.523, indicating that forest cover loss correlates with LST increase. Regression analysis suggests that each 1 km² increase in forest cover is associated with a 0.028 °C decrease in LST. These findings underscore the critical role of land cover management in mitigating urban warming.

The segmentation and classification of high-resolution satellite images (HRSI) are useful approaches to extract information. In recent times, roads and buildings have been classified for analysis of urban areas in a better manner. Apart from these, healthy trees are also an important factor in HRSI, i.e. adjacent to roads, and vegetation. They reflect the area in an image as land cover. Other important information, shadow, is extracted from satellite images, which indicates the presence of trees and built-up areas such as buildings, flyovers, etc. In this article, a weighted membership-function-based fuzzy c-means with spatial constraints (WMFCSC) approach for automated satellite image classification is proposed. Initially, spatially fuzzy clustering is used to classify the satellite images in healthy trees with vegetation, roads, and shadows, which includes the information of spatial constraints. The road results of the classified image are still having non-road segments. Therefore, the proposed four intermediate stages (IS) are used to extract the road information, followed by the results of road areas of the WMFCSC approach. The framework of IS helps to remove the false road segments which are adjacent to roads and renovates the segmented roads due to the shadow effect. A final step of a hybrid WMFCSC-IS approach is used to extract the road network. The results of classified images confirm the effectiveness of the WMFCSC-IS approach for satellite image classification.

Trait predictions from leaf spectral properties are mainly applied to tree species, while herbaceous systems received little attention in this topic. Whether similar trait–spectrum relations can be derived for herbaceous plants that differ strongly in growing strategy and environmental constraints is therefore unknown. We used partial least squares regression to relate key traits to leaf spectra (reflectance, transmittance, and absorbance) for 35 herbaceous species, sampled from a wide range of environmental conditions. Specific Leaf Area and nutrient-related traits (N and P content) were poorly predicted from any spectrum, although N prediction improved when expressed on a per area basis (mg/m2 leaf surface) instead of mass basis (mg/g dry matter). Leaf dry matter content was moderately to good correlated with spectra. We explain our results by the range of environmental constraints encountered by herbaceous species; both N and P limitations as well as a range of light and water availabilities occurred. This weakened the relation between the measured response traits and the leaf constituents that are truly responsible for leaf spectral behavior. Indeed, N predictions improve considering solely upper or under canopy species. Therefore, trait predictions in herbaceous systems should focus on traits relating to dry matter content and the true, underlying drivers of spectral properties.

Data fusion techniques are currently used to merge low resolution multispectral with high resolution panchromatic data. Various authors have expended efforts on attempting to achieve a fusion where the spectral characteristics of one image are combined with the spatial attributes of the second producing a new, fused image. This philosophy has a number of inherent problems. The first is that of retaining the spectral fidelity of the original hyperspectral data in the fused data set. Techniques currently used to achieve the fusion usually do not retain the spectral characteristics accurately. Associated with this is the assumption that the intensity of distributions associated with the high spatial resolution (often panchromatic) image, usually collected in the visible portion of the spectrum, are consistent throughout the entire spectral range in question. Another problem is that of image size. To fuse a 20 meter MSS with a 1 meter panchromatic image represents an associated 400 fold increase in data base size. The approach in this extracts the spatial attributes of the high resolution data set (in our case the individual tree crowns). Once extracted, the spatial properties of the extracted attributes are assessed and translated into a spatial coverage consistent with the resolution of the spectral data for further analyses. The advantages of this approach are firstly that the data are used at the resolutions that they were originally collected. Secondly, it avoids the creation of large, unmanageable images.

Hundreds of narrow contiguous spectral bands collected by a hyperspectral sensor have provided the opportunity to identify the various materials present on the surface. Moreover, spatial information, enforcing the assumption that the adjacent pixels belong to the same class with a high probability, is a valuable complement to the spectral information. In this paper, two predominant approaches have been developed to exploit the spatial information. First, by decomposing each pixel and the spatial neighborhood into a low-rank form, the spatial information can be efficiently integrated into the spectral signatures. Meanwhile, in order to describe the low-rank structure of the decomposed data more precisely, an $\ell_{1/2}$ norm regularization is introduced and a discrete algorithm is proposed to solve the combined optimization problem by the augmented Lagrange multiplier (ALM) and a half-threshold operator. Second, a graph cuts segmentation algorithm has been applied on the sparse-representation-based probability estimates of the hyperspectral data to further improve the spatial homogeneity of the material distribution. Experimental results on four real hyperspectral data with different spectral and spatial resolutions have demonstrated the effectiveness and versatility of the proposed spatial information-fused approaches for hyperspectral image classification.

Detailed land cover information is valuable for mapping complex urban environments. Recent enhancements to satellite sensor technology promise fit-for-purpose data, particularly when processed using contemporary classification approaches. We evaluate this promise by comparing the influence of spatial resolution, spectral band set and classification approach for mapping detailed urban land cover in Nottingham, UK. A WorldView-2 image provides the basis for a set of 12 images with varying spatial and spectral characteristics, and these are classified using three different approaches (maximum likelihood (ML), support vector machine (SVM) and object-based image analysis (OBIA)) to yield 36 output land cover maps. Classification accuracy is evaluated independently and McNemar tests are conducted between all paired outputs (630 pairs in total) to determine which classifications are significantly different. Overall accuracy varied between 35% for ML classification of 30 m spatial resolution, 4-band imagery and 91% for OBIA classification of 2 m spatial resolution, 8-band imagery. The results demonstrate that spatial resolution is clearly the most influential factor when mapping complex urban environments, and modern “very high resolution” or VHR sensors offer great advantage here. However, the advanced spectral capabilities provided by some recent sensors, coupled with contemporary classification approaches (especially SVMs and OBIA), can also lead to significant gains in mapping accuracy. Ongoing development in instrumentation and methodology offer huge potential here and imply that urban mapping opportunities will continue to grow.

Markov random field (MRF) based methods have been widely used in high spatial resolution (HSR) image classification. However, many existing MRF-based methods put more emphasis on pixel level contexts while less on superpixel level contextual information. To cope with this issue, this article presents a novel bilevel contextual MRF framework, named BLC-MRF, for HSR imagery classification. Specifically, pixel and superpixel level dependence are incorporated into the proposed MRF model to fully exploit spectral–spatial contextual information and preserve object boundaries in HSR images. In BLC-MRF, a pixel level MRF model is first performed and then cascaded as an input of a superpixel level MRF. In superpixel level, unary and pairwise potential terms are constructed by using the superpixel probability estimation method and spectral histogram distance, respectively. At last, a contextual MRF model is conducted and the final classification map can be computed by using $\alpha$ -expansion algorithm. The benefits of BLC-MRF are twofold: first, the pixel and superpixel level contextual information can be exploited under MRF framework to preserve object boundaries for improving the classification performance, and, second, the algorithm can provide promising results with a small number of training samples. Experimental results on three HSR datasets demonstrate that the proposed approach outperforms several state-of-the-art methods in terms of the classification performance.

ABSTRACTRemote sensing technology can be a valuable tool for mapping coral reef ecosystems. However, the resolution capabilities of remote sensors, the diversity and complexity of coral reef ecosystems, and the low reflectivity of marine environments increase the difficulties in identifying and classifying their features. This research study explores the capability of high spatial resolution (WorldView-2 (WV-2) and Pleiades-1B) and low spatial resolution (Land Remote-Sensing Satellite (Landsat 8)) multispectral (MS) satellite sensors in quantitatively mapping coral density. The Kubbar coral reef ecosystem, located in Kuwait’s southern waters, was selected as the research site. The MS imagery of WV-2, Pleiades-1B and Landsat 8 were, after geometric and radiometric assessment and corrections, subjected to new image classification approach using a Multiple Linear Regression (MLR) analysis. The new approach of MLR coral density analysis used the dependent variable of coral density percentage from ground truth and independent variables of spectral reflectance from selected imagery, depth (as estimated from a surface derived from bathymetric charts) and distance to land or reef unit centre. Accuracy assessment using independent ground truth was performed for the selected approach and satellite sensors to determine the quality of the information derived from image classification processes. The results showed that coral density maps developed using the MLR coral density model proved to have some level of reliability (radiometrically corrected WV-2 image (the coefficient determination denoted as R-squared (R²) = 0.5, Root-Mean-Square Error (RMSE) = 10) and radiometrically corrected Pleiades-1B image (R² = 0.8, RMSE = 10)). This study suggested using high spectral resolution data and including additional factors (variables) (e.g. water turbidity, temperature and salinity) could contribute to improving the accuracy of coral density maps produced by application of the MLR model; however, all of these would add cost and effort to the mapping process. The outcomes of this research study provide coral reef ecosystem researchers, managers, and decision makers a tool to determine and map coral reef density in more detail than in the past. It will help quantify coral density at particular points in time leading to estimates of change, and allow coral reef ecologists to identify the current coral reef habitat health status, distribution and extent.

In this paper, we present a novel spatial and spectral fusion model (SASFM) that uses sparse matrix factorization to fuse remote sensing imagery with different spatial and spectral properties. By combining the spectral information from sensors with low spatial resolution (LSaR) but high spectral resolution (HSeR) (hereafter called HSeR sensors), with the spatial information from sensors with high spatial resolution (HSaR) but low spectral resolution (LSeR) (hereafter called HSaR sensors), the SASFM can generate synthetic remote sensing data with both HSaR and HSeR. Given two reasonable assumptions, the proposed model can integrate the LSaR and HSaR data via two stages. In the first stage, the model learns from the LSaR data a spectral dictionary containing pure signatures, and in the second stage, the desired HSaR and HSeR data are predicted using the learned spectral dictionary and the known HSaR data. The SASFM is tested with both simulated data and actual Landsat 7 Enhanced Thematic Mapper Plus (ETM+) and Terra Moderate Resolution Imaging Spectroradiometer (MODIS) acquisitions, and it is also compared to other representative algorithms. The experimental results demonstrate that the SASFM outperforms other algorithms in generating fused imagery with both the well-preserved spectral properties of MODIS and the spatial properties of ETM+. Generated imagery with simultaneous HSaR and HSeR opens new avenues for applications of MODIS and ETM+.

New high-resolution sensors offer the potential to apply remote sensing to a new, finer scale of problems. However, the highest spatial resolution imagery typically has only a single spectral band, and therefore, classification based on spatial patterns is required. One application particularly suited to spatial classification is the identification of orchards and vineyards. Orchards and vineyards have a distinctive repeating pattern that can be identified by an analysis of local autocorrelation patterns. A classification algorithm based on an analysis of autocorrelograms was developed, and tested using Ikonos panchromatic imagery of Granger, Washington. The spatial autocorrelation-based classification resulted in an estimated accuracy of 0.954 and � of 0.900. Errors of omission for orchards and vineyards were slightly higher than errors of commission (11‐14 percent versus 3‐6 percent). By comparison, a maximum likelihood classification of 32 gray level co-occurrence texture