Contextual classification of multispectral pixel data

doi:10.1016/0262-8856(84)90040-4

Image and Vision Computing

Volume 2, Issue 1, February 1984, Pages 13-29

https://doi.org/10.1016/0262-8856(84)90040-4 Get rights and content

Abstract

Contextual statistical decision rules for classification of lattice-structured data such as pixels in multispectral imagery are developed. Their recursive implementation is shown to have a strong resemblance to relaxation algorithms. Experimental evaluation of the proposed algorithms demonstrates their effectiveness.

References (39)

GT Toussaint
The use of context in pattern recognition
Pattern Recognition
(1978)
S Wharton
A contextual classification method for recognising land use patterns in high resolution remotely sensed data
Pattern Recognition
(1982)
DA Landgrebe
The development of a spectral-spatial classifier for earth observational data
Pattern Recognition
(1980)
N Ahuja et al.
Neighbor gray levels as features in pattern classification
Pattern Recognition
(1980)
BR Hunt
Nonstationary statistical image models (and their application to image data compression)
Comput. Graphics Image Process
(1980)
RL Kashyap et al.
Decision rules for choice of neighbors in random field models of images
Comput. Graphics Image Process
(1981)
PH Swain et al.
Contextual classification of multispectral data
Pattern Recognition
(1981)
TS Yu et al.
Recursive contextual classification using a spatial stochastic model
Pattern Recognition
(1983)
KS Fu et al.
Statistical pattern classification using contextual information
(1980)
A Rosenfeld et al.
Digital picture processing
(1982)

KI Itten et al.

Thematic adaptive spatial filtering of landsat land use classification results

JM Brayer et al.

Modelling of earth resources satellite data

KS Fu

Syntactic (linguistic) pattern recognition

DC Mason

Segmentation of terrain images using textural and spectral characteristics

Comput. Digital Techniques

(1979)

T Pavlidis et al.

Segmentation by texture using a co-occurrence matrix and a split and merge algorithm

Comput. Graphics Image Process

(1979)

J Desachy

Texture features in remote sensing imagery

RM Haralick et al.

Textural features for image classification

IEEE Trans. Syst., Man Cybern.

(1973)

JS Weszka et al.

A comparative study of texture measures for terrain classification

IEEE Trans. Syst., Man Cybern.

(1976)

GR Cross et al.

Markov random field texture models

IEEE Trans. Pattern Anal. Mach. Intell.

(1983)

Cited by (80)

Optimizing wind farms layouts for maximum energy production using probabilistic inference: Benchmarking reveals superior computational efficiency and scalability
2021, Energy
Citation Excerpt :
Note that other inference algorithms exist and could be used to find low-energy solutions. For example, gradient descent is considered one of the simplest optimization techniques, which minimizes the energy by changing independent subsets of variables to find lower-energy configurations [67]. Gradient descent however tends to get stuck in local minima, and stochastic gradient descent are typically implemented to overcome this limitation.
Successful development of wind farms relies on the optimal siting of wind turbines to maximize the power capacity under stochastic wind conditions and wake losses caused by neighboring turbines. This paper presents a novel method to quickly generate approximate optimal layouts to support infrastructure design decisions. We model the quadratic integer formulation of the discretized layout design problem with an undirected graph that succinctly captures the spatial dependencies of the design parameters caused by wake interactions. On the undirected graph, we apply probabilistic inference using sequential tree-reweighted message passing to approximate turbine siting. We assess the effectiveness of our method by benchmarking against a state-of-the-art branch and cut algorithm under varying wind regime complexities and wind farm discretization resolutions. For low resolutions, probabilistic inference can produce optimal or nearly optimal turbine layouts that are within 3% of the power capacity of the optimal layouts achieved by state-of-the-art formulations, at a fraction of the computational cost. As the discretization resolution (and thus the problem size) increases, probabilistic inference produces optimal layouts with up to 9% more power capacity than the best state-of-the-art solutions at a much lower computational cost.
A new approach to contextual learning using interval arithmetic and its applications for land-use classification
2016, Pattern Recognition Letters
Citation Excerpt :
When looking at a picture or a video, we can clearly see the pixels vary smoothly in homogeneous regions. Sequential- and contextual-oriented learning are some well-known methodologies very often used to address situations in which spatial and/or contextual information may help the classifier into better modeling the behavior of the data, as stated by Cohen and Carvalho [2], which introduced the Stacked Sequential Learning (SSL), as well as by Kittler and Föglein] [9] and Dietterich [3], that presented an interesting review about sequential learning techniques. The authors also highlighted the high computational load of some techniques based on such idea.
Contextual-based classification has been paramount in the last years, since spatial and temporal information play an important role during the process of learning the behavior of the data. Sequential learning is also often employed in this context in order to augment the feature vector of a given sample with information about its neighborhood. However, most part of works describe the samples using features obtained through standard arithmetic tools, which may not reflect the data as a whole. In this work, we introduced the Interval Arithmetic to the context of land-use classification in satellite images by describing a given sample and its neighbors using interval of values, thus allowing a better representation of the model. Experiments over four satellite images using two distinct supervised classifiers showed we can considerably improve sequential learning-oriented pattern classification using concepts from Interval Arithmetic.
Improving land cover classification through contextual-based optimum-path forest
2015, Information Sciences
Citation Excerpt :
Moser and Serpico [27] also proposed a version of the SVM–MRF approach using a single-step formulation, rather than the two-phase strategy presented in [40]. Kittler and Föglein [20] combined a Bayesian classifier and a Gaussian MRF by using pixel spectral and spatial dependencies. Shekhar et al. [38] employed a Bayesian classifier with different models to capture spatial context.
Traditional machine learning algorithms very often assume statistically independent data samples. However, this is clearly not the case in remote sensing image applications, in which pixels present spatial and/or temporal dependencies. In this work, it has been presented an approach to improve land cover image classification using a contextual approach based on optimum-path forest (OPF) and the well-known Markov random fields (MRFs), hereinafter called OPF–MRF. In addition, it is also introduced a framework to the optimization of the amount of contextual information used by OPF–MRF. Experiments over high- and medium-resolution satellite (CBERS-2B, Landsat 5 TM, Ikonos-2 MS and Geoeye) and radar (ALOS-PALSAR) images covering the area of two Brazilian cities have shown the proposed approach can overcome several shortcomings related to standard OPF classification. In some cases, the proposed approach outperformed traditional OPF in about 9% of recognition rate, which is crucial for land cover classification.
A probabilistic framework for next best view estimation in a cluttered environment
2014, Journal of Visual Communication and Image Representation
In this article, we present an information gain-based variant of the next best view problem for occluded environment. Our proposed method utilizes a belief model of the unobserved space to estimate the expected information gain of each possible viewpoint. More precise, this belief model allows a more precise estimation of the visibility of occluded space and with that a more accurate prediction of the potential information gain of new viewing positions. We present experimental evaluation on a robotic platform for active data acquisition, however due to the generality of our approach it also applies to a wide variety of 3D reconstruction problems. With the evaluation done in simulation and on a real robotic platform, exploring and acquiring data from different environments we demonstrate the generality and usefulness of our approach for next best view estimation and autonomous data acquisition.
Fast deep mixtures of Gaussian process experts
2024, Machine Learning
A machine learning route between band mapping and band structure
2023, Nature Computational Science

View all citing articles on Scopus

View full text

Contextual classification of multispectral pixel data

Abstract

Pattern Recognition

Pattern Recognition

Pattern Recognition

Pattern Recognition

Comput. Graphics Image Process

Comput. Graphics Image Process

Pattern Recognition

Pattern Recognition

Statistical pattern classification using contextual information

Digital picture processing

Thematic adaptive spatial filtering of landsat land use classification results

Modelling of earth resources satellite data

Syntactic (linguistic) pattern recognition

Segmentation of terrain images using textural and spectral characteristics

Comput. Digital Techniques

Segmentation by texture using a co-occurrence matrix and a split and merge algorithm

Comput. Graphics Image Process

Texture features in remote sensing imagery

Textural features for image classification

IEEE Trans. Syst., Man Cybern.

A comparative study of texture measures for terrain classification

IEEE Trans. Syst., Man Cybern.

Markov random field texture models

IEEE Trans. Pattern Anal. Mach. Intell.