Evaluate the accuracy of image classification algorithms in the integration of spatial-temporal satellite images
Evaluate the accuracy of image classification algorithms in the integration of spatial-temporal satellite images
Ali Rakhsh Mahpour1 Rouzbeh Shad2 A. Reza Faridhosseini3
1) Iran Ferdowsi University of Mashhad
2) Iran Ferdosi University of Mashhad
3) Iran Ferdosi University of Mashhad
IOI Code
:
XYBB-ZHDBE
Publication :
3rd International Conference On Research Science And Technology(3rstconf.com)
Abstract :
Todays, the use of satellite imagery as a tool for monitoring different phenomena. However, satellites images are available on a daily, have low spatial resolution. therefore The daily phenomenon cannot be accurately monitored by high spatial resolution. But in recent years the development of algorithms for spatial-temporal fusion of satellite imagery has made it possible. STARFM is one of the most famous spatial-temporal fusion algorithms. appropriate image classification is the most important initial steps of this algorithms. Therefore, in this study, the accuracy of the image classification algorithm on STARFM performance, by integration of Landsat and the Modis images, is studied. The results indicate that the accuracy of the preprocessing and other errors directly effected on the algorithm results and the Uniformity of study area Increases the accuracy of the final results. Although, Uniformity and homogeneity factor is much higher impact than the accuracy of image classification. in The non-uniform areas, increasing the kappa coefficient, will not improve the accuracy of the algorithm Necessarily. As well as other error factors (e.g. non-same spectral range) will make a significant impact on results.
Keywords : fusion temporal spatial image classification satellite image