Pattern Recognition Principles
Taşkın Kavzoğlu, Brandt Tso, Paul M. Mather · CRC Press eBooks · 2024
Remote sensing classification that deals with high-dimensional data with complex spectral and radiometric characteristics can be achieved with pattern recognition techniques. Remotely sensed image data exhibit a high level of complexity and irregularity so that pixels within these images frequently contain multiple features, a composition influenced by the spatial resolution of the sensor and the configuration of the surveyed area. Therefore, the classification of remotely sensed data into a thematic map continues to be a challenge because many factors may have significant effects on the success of a classification. Recognizing patterns allows us to make predictions about unseen pixels or objects. In the context of pattern recognition, a pattern is a vector of features describing an object. This pattern is made up of measurements on a set of features, which can be thought of as the axes of a k -dimensional space, called the feature space. The aim of pattern recognition is to establish a relationship between a pattern and a class label. This relationship may be one-to-one (producing a hard classification) or one-to-many (producing a fuzzy classification). The features describing the object may be spectral reflectance or emittance values from optical or infrared imagery, radar backscatter values, secondary measurements derived from the image, called ancillary data, such as texture, or geographical features such as terrain elevation, slope, and aspect. This chapter introduces the principles of pattern recognition, starting from the terminology used in remote sensing image classification and the taxonomy of classification techniques. Details of the well-known supervised and unsupervised statistical classifiers are then described. Ensemble classifiers and the incorporation of secondary information are also discussed in later chapters.
Cited by 1,053 works. doi:10.1201/9781003439172-2