A Comparative Study of Supervised Classification Techniques for Multi-Spectral Images | ||||
The International Conference on Electrical Engineering | ||||
Article 39, Volume 11, 11th International Conference on Electrical Engineering ICEENG 2018, April 2018, Page 1-13 PDF (588.86 K) | ||||
Document Type: Original Article | ||||
DOI: 10.21608/iceeng.2018.30172 | ||||
View on SCiNiTO | ||||
Authors | ||||
Mahmoud abdallah Shwaky; Fawzy Eltohamy Hassan Amer; Osama M. Mosa; Essam Hamza | ||||
Egyptian Armed Forces. | ||||
Abstract | ||||
Classification of satellite images is an important key for ground features extraction and thematic maps production. Satellite images with multi-spectral bands provide rich data which is useful for features extraction and description. Many supervised classification methods have been developed for classifying the multispectral images. Each method has its own advantages and disadvantages (limitations). In this paper the performance of four of the common used supervised classification techniques is compared. The techniques considered here are: Parallelepiped (PP), Minimum Distance (MD), Mahalanobies (MA), and Maximum Likelihood (ML). They are applied on a set of multispectral images acquired by Worldview-2 satellite. The classification results accuracy are analyzed and evaluated The research work flow is processed by using ENVI. The developed maps are then visually compared with each other and accuracy assessments utilizing ground-truths. The assessment of classification results is represented in confusion matrix format and determination of Kappa coefficients. The preliminary results show that Maximum Likelihood (ML) gives accurate classification result for the area of study with overall accuracy 91.5741% and it is evaluated by Kappa coefficient which is 0.8846: | ||||
Keywords | ||||
supervised classification methods; Image classification assessment | ||||
Statistics Article View: 280 PDF Download: 342 |
||||