Satellite Image Classification

Machine learning algorithms also known as artificial intelligence is used in variety of fields to help mankind. A robot can be made intelligent by making it recognize diseased plant leaves and also recognize objects that can be used in industries for sorting and classification making less use of manpower and saving money.

The satellite images obtained from specific satellites processed and utilised to make maps for urban planning and also identify critical resources available on the earth without field inspection. The landslides area, forest fire areas all can be inspected from the satellite images. But Raw satellite image is not that good to identify all types of land covers.

Classification algorithms group the pixels of similar characteristics and a proper better readable and efficient maps can be obtained from raw satellite data.

Fig.1 Raw satellite data                                                            Fig. 2: Classified Satellite image using
                                                                                                             Fuzzy C-Means algorithm

Six Land Cover Classes have been identifed from the input image by FCM algorithm.

The Fuzzy C-Means is an machine learning algorithm that classify the pixels of the input image without training. All the pixels of the input image are checked to which land Cover class it belongs. The pixels are placed to specific land cover class according to its belonging using distance measure techniques. Therefore pixels of the same Land Cover are grouped and better image is displayed as output.


Thanks for Reading