Classification Of Multisensor Images Research Articles

Assessment of glacier retreat in the Sierra Nevada del Cocuy, Colombia based on multisensor image classification

Los glaciares andinos representan una de las fuentes principales del recurso hídrico en Suramérica y durante las últimas décadas se han reducido significativamente como producto del cambio climático y la variabilidad climática. En los Andes colombianos, el pico nevado más extenso corresponde a la Sierra Nevada del Cocuy (SRC), un cordón montañoso localizado al noreste de la Cordillera Oriental con presencia de nieves perpetuas en alturas que oscilan aproximadamente entre los 4800 y los 5345 metros sobre el nivel del mar (msnm). A partir de imágenes satelitales de Landsat-4 (1987), Landsat-5 (1991, 1997, 2009), Landsat-7 (2000, 2003), Landsat-8 (2014, 2016, 2017), y Sentinel-2 (2019, 2021) se realizó una clasificación orientada a píxel usando el software PCI Geomatics, en la cual se definieron 4 tipos de cobertura: área glaciar, suelo-roca, vegetación y agua. Para la validación de exactitud (accuracy) fueron utilizadas como datos de referencia, imágenes satelitales de alta resolución espacial (Google Earth ~1,0 m y Planet’s high-resolution, analysis-ready mosaics of the world’s tropics ~4,7 m) y puntos de control de campo. Los valores de exactitud global (todas las coberturas) oscilaron entre 86-99%, con una exactitud para la cobertura de área glaciar entre 97-100%. La disminución de dicha área es de 1099,59 ha en un lapso de 34 años (1987-2021). Este análisis reveló que el área glaciar disminuyó aproximadamente en un 37,92% con respecto a la primera escena (1987). Según dicha tendencia, el glaciar de la SRC se extinguiría para el año 2048. La tasa de retroceso glaciar está influenciada principalmente por factores relacionados con el calentamiento global como lo son el aumento de la temperatura media anual y la disminución en las tasas de precipitación, y factores de variabilidad climática como el fenómeno de El Niño.

Classification of Multisensor Images with Different Spatial Resolution

The paper is focused on the analysis of classification possibilities of multisensor data with different spatial resolutions using combined classifiers based on Bayes approach with equal prior probabilities and on minimum of the Mahalanobis distance. The task set up for the 2014 IEEE GRSS Data Fusion Contest was chosen as an application example. High resolution RGB image and lower resolution thermal infrared image from the same urban area were processed to perform classification of each higher resolution pixel. Development of a fast and straightforward procedure was targeted and combined classifiers are proposed for that, exploiting spectral features from each data set separately. It is shown that data fusion can be achieved using the proposed classifiers and improvement of classification quality can be obtained with respect to the cases where only one of the data sets is used. The best classification results were obtained using the combined Bayes- type classifier that provided overall classification accuracy of about 95 % when the ground truth pixels from the high resolution RGB image were used both for design and testing. DOI: http://dx.doi.org/10.5755/j01.eee.21.5.13333

Classification of Multisensor Images with Different Spatial Resolution

The paper is focused on the analysis of classification possibilities of multisensor data with different spatial resolutions using combined classifiers based on Bayes approach with equal prior probabilities and on minimum of the Mahalanobis distance. The task set up for the 2014 IEEE GRSS Data Fusion Contest was chosen as an application example. High resolution RGB image and lower resolution thermal infrared image from the same urban area were processed to perform classification of each higher resolution pixel. Development of a fast and straightforward procedure was targeted and combined classifiers are proposed for that, exploiting spectral features from each data set separately. It is shown that data fusion can be achieved using the proposed classifiers and improvement of classification quality can be obtained with respect to the cases where only one of the data sets is used. The best classification results were obtained using the combined Bayes- type classifier that provided overall classification accuracy of about 95 % when the ground truth pixels from the high resolution RGB image were used both for design and testing.DOI: http://dx.doi.org/10.5755/j01.eee.21.5.13333

Segment-based Image Classification of Multisensor Images

This study proposed two multisensor fusion methods for segment-based image classification utilizing a region-growing segmentation. The proposed algorithms employ a Gaussian-PDF measure and an evidential measure respectively. In remote sensing application, segment-based approaches are used to extract more explicit information on spatial structure compared to pixel-based methods. Data from a single sensor may be insufficient to provide accurate description of a ground scene in image classification. Due to the redundant and complementary nature of multisensor data, a combination of information from multiple sensors can make reduce classification error rate. The Gaussian-PDF method defines a regional measure as the PDF average of pixels belonging to the region, and assigns a region into a class associated with the maximum of regional measure. The evidential fusion method uses two measures of plausibility and belief, which are derived from a mass function of the Beta distribution for the basic probability assignment of every hypothesis about region classes. The proposed methods were applied to the SPOT XS and ENVISAT data, which were acquired over Iksan area of of Korean peninsula. The experiment results showed that the segment-based method of evidential measure is greatly effective on improving the classification via multisensor fusion.

A multigradient algorithm using a mixture of experts architecture for land cover classification of multisensor images

An algorithm for supervised classification of multisensor images is proposed. The mixture of experts (ME) architecture with dynamic weight allocation is used for multiclass classification. Here the classification is treated as a maximum likelihood problem and the synaptic weights of the expert network and gating network are updated by a stochastic multigradient approach. Data from an optical sensor with four bands and a synthetic aperture radar (SAR) image of the same scene has been fused and classified. The algorithm is compared to some other advanced training algorithms in the literature for the same image data.

Pruning artificial neural networks: An example using land cover classification of multi-sensor images

The use of three techniques for pruning artificial neural networks (magnitude-based pruning, optimum brain damage and optimal brain surgeon) is investigated, using microwave SAR and optical SPOT data to classify land cover in a test area located in eastern England. Results show that it is possible to reduce network size significantly without compromising overall classification accuracy; indeed, accuracy may rise as the number of links decreases. However, individual class accuracies and the spatial distribution of the pixels forming the individual classes may change significantly. If the network is pruned too severely some classes may be eliminated altogether. In terms of maintaining overall classification accuracy the optimal brain surgeon algorithm gave the best results, and magnitude-based pruning also gave good results despite its simplicity. The optimum brain damage algorithm performed least well of the three methods tested.