Extraction of image objects in very high resolution satellite images using spectral behaviour in look up table and color space based approach
Pankaj Pratap Singh · 2016 SAI Computing Conference (SAI) · 2016
In this work, a hybrid approach is proposed to extract the image objects from high resolution satellite images (HRSI). The extraction of image objects is a one kind of incubation process in HRSI that is still an interesting field. It is also a challenging area due to the presence of shadow effect in the images which can reduce the effectiveness of image classification. To overcome this shadow effect area for better image classification, Look-up table (LUT) is used for assigning the index values corresponding the intensity of the spectral behaviour of satellite images. With the help of LUT, intensity values of different object are evaluated and stored which are used to distinguish the objects in this approach. Furthermore, after the analysis of the different kind of LUT's those are having the color intensities. In this regarding, non-linear function based color space method is used in HRSI that can help to susceptible the objects of similar pixel values. Shadow effects can occluded the correct information of extracted roads and such kind of issue is resolved with this proposed hybrid approach.