Comparative Analysis of Pixel-Based and Object-Based Classification of High Resolution Remote Sensing Images – A Review

We delineate an overall performance comparison between the two most popular classification techniques: Pixel-Based and ObjectBased of remote sensing. Object based image analysis has been widely used as a common paradigm in the analysis of high resolution remotely sensed satellite data which is used to extract the meaningful information for updating the GIS data. Both techniques have their own pros and cons in finding the solutions of many applications. To create land cover thematic maps with higher accuracies then the image classification analysis is a big challenge in the remote sensing. However, the pixel based classification technique works only on spectral features and neglects the spatial features but in object based classification have ability to work on both spectral and spatial features. OBC has also characteristic features like mean, standard deviation etc. which can be used to differentiate the classes properly. In paper, we observed that the object-based technique shows higher accuracy in classification process than the pixel-based technique because pixel based can’t satisfy the high resolution satellite data properties and it produced data redundancy.