Multispectral image classification

Multi spectral images are the imagery that are commonly obtained through the satellites. These images are called multi-spectral since these images are acquired for the same area for different wavelengths of the electromagnetic energy. For example Landsat satellite, Multi spectral scanner - band 1 image, band 2 image etc. Since these remote sensing images are typically multispectral responses of various features it is hard to identify directly the feature type by visual inspection. Hence the remote sensing data has to be classified first then it is processed using various data enhancement techniques so as to give the user an idea about the features that are present in the image.

And this classification of the remote sensing images is a complex task which involves rigorous validation of the training samples depending on the classification algorithm used for the classification of the image. The techniques that are used for the classification of Remote sensing data can be grouped mainly into two types.

Supervised classification techniques
Unsupervised classification techniques

Supervised Classification Techniques

Supervised classification makes use of the training samples. Training samples are nothing but the areas on the ground that we already know about the features that are present above it. Using the spectral signatures of the training areas we will search for the similar type of spectral signatures in the remaining pixels of the image, and we will classify accordingly. This type of classification which uses the training samples for classification is called supervised classification. And expert knowledge is very important in this type of classification since the selection of the training samples and adopting a bias can badly affect the accuracy of classification. one of the popular technique that is widely used in this type of classification is Maximum Liklihood principle. In this we will calculate the probability of a pixel belonging to a class i.e., feature and will allot the pixel to class with maximum probability,

Unsupervised Classification

In case of unsupervised classification no priory knowledge is required for classifying the features of the image. In this the natural clustering or grouping of the pixel values i.e., gray levels of the pixels are observed. Then a threshold level is defined for adopting the no of classes in the image. finer the threshold value more will be the no of classes. But beyond a certain limit same class is represented in different classes in the sense variation in the class is represented. After forming the clusters, ground truth validation is done to identify the class the image pixel belongs to. Thus in this unsupervised classification apriori information about the classes is not required. One of the popular method in unsupervised classification is K means classifier algorithm.

References