Investigating the Possibility of Feature Extraction in Subpixel Scale from MODIS Images in Homogeneous Backgrounds

Document Type : مقاله پژوهشی


Master of Remote Sensing and GIS


Today, due to the access importance to precise, comprehensive and daily spatial information of the various phenomena and the change over time in terms of location and nature. In order to monitor their behavior and the possibility of planning and deciding in different cases, it have been attempting to improve the imaging performance with increasing spatial accuracy, radiometric spectra of the sensors. But each improvement stage requires time and cost, and taking into account different periods and sometimes long periods of imaging of a point on Ground level (low time resolution), which is due to smaller FOV (increased spatial accuracy). However, daily monitoring of many features is not possible, especially those that are smaller than the sensor's pixel size. Therefore, it is necessary to use, methods for features extraction from low resolution images, in which the feature dimensions and the spatial resolution of sensor is of fundamental situation. In this research, attempt has been made to investigate the possibility of various possible modes, including the location of the barge in the image, the images taken in different bands of emission, radiance, reflection, the presence of two features side by side, the effect of flame on adjacent feature, the study of noisy bands, the presence or absence of flame, non-solid features such as platforms and solid features, such as barge, on different days of the year.