Application of Fuzzy Logic and Neural Network in Crop Classification: A Review

Abstract

Image classification is one of the crucial techniques in detecting the crops from remotely sensed data as mapping of crops is a complex activity which in turn is an important parameter for planning and management of irrigation command area. Classifying remotely sensed data into a thematic map remains a challenge because many factors, such as the complexity of the landscape in a study area, selected remotely sensed data, and image-processing and classification approaches, may affect the success of a classification. Up-to-date and accurate classification results are required for analyses, which provide basis for deciding and implementing policies and plans for management of agricultural crops in local, regional and global scale. Successful identification of crops requires knowledge of the developmental stages and appearance of each crop in the area to be inventoried. The crops thus identified from the remote sensing data can be utilized to estimate crop water requirements for irrigation planning and management. Supervised and unsupervised classifications have been in common use in remote sensing for many years which are known as hard classification. Supervised and unsupervised classifications rely on classical set theory in assigning pixels into discrete classes based on training sets and some statistically determined criteria. Conventional classification methods such as Supervised and Unsupervised classification techniques are often incapable of performing satisfactorily in the presence of mixed pixels, the pixels which are not completely occupied by a single homogeneous category. Soft computing techniques are useful for tackling these real-world problems based on fuzzy systems, artificial neural networks, and evolutionary algorithms. These techniques are widely used nowadays by researchers. Nevertheless, each model has its own advantages and disadvantages.Contrary to hard classifiers, fuzzy classifier does not make a definitive decision about the land cover class to which each pixel belongs. Rather, they develop statements of the degree to which each pixel belongs to each of the land cover classes being considered. It is achieved by applying a function called "membership function" on remotely sensed images. Artificial Neural Networks are computer programs that are designed to simulate human-learning processes through establishment and reinforcement of linkages between input data and output data. Artificial neural network is used as a powerful tool for pattern classification and have been found to be accurate in the classification of remotely sensed data. different soft computing classification techniques for crop mapping which is necessary for estimating crop water requirements with the help of satellite images.

Hyperspectral Image Classification using Softcomputing Techniques: A Review

“…Per pixel classifiers may be parametric or non parametric. These methods can be used in decision trees and Support Vector Machine [17].…”

Fuzzy logika az agrár- és környezetkutatásokban

“…MURMU és BISWAS (2015) rámutat arra, hogy a hagyományos képértékelésnél a vegyes pixelek nem a megfelelő kategóriába kerülnek, vagy értékelésük elmarad, mivel ezek nem sorolhatóak egyetlen homogén kategóriába sem. Véleményük szerint e problémára megoldást jelenthetnek a fuzzy rendszerek, amennyiben az egyes pixeleket tagsági függvények segítségével aszerint értékeljük, hogy mekkora mértékben esnek az egyes kategóriákba.…”

Fluorescence and Reflectance Spectroscopy for Early Detection of Different Mycorrhized Plantain Plants

“…Support Vector Machine are important advanced techniques intensively used (Mountrakis & Ogole, 2011;Moshou, Pantazi, Kateris, & Gravalos, 2014). Others efficient tools are concerned with fuzzy logic, neural network and extreme learning machine (Moreno, Corona, Lendass, Graña, & Galvão 2014;Murmu & Biswas, 2015;Cvetković, Stojanović, & Nicolić, 2015).The boosting meta-algorithm is relatively new, efficient, simple, and easy to manipulate additive modeling technique that can use potentially any weak learner available. LogitBoost is another variant of the boosting algorithm that performs additive logistic regression (Friedman, Hastie, & Tibshirani, 2000) and has shown better performance than many other machine learning algorithm especially in protein data structure classification (Krishnaraj & Reddy, 2009).…”