ASSESSMENT OF CLOUD AREA USING FIELD SERVER AND IMAGE PROCESSING

摘要

Knowledge regarding the cloudiness is very important in Metrological Science. Cloud area is one significant factor to indicate the weather condition. Weather is an external factor to influence crop growth. In this work, the field server with common camera which is used to observe both a crop field and the weather at the same time is proposed. The main purpose of this research is to automatically assess cloud area in the sky from the images that are provided by field server. The image is a composition of the field of the crop such as rice, sugarcane and cassava etc. for monitoring crop growth and the sky area for observing the weather condition. In this research, the sky area is only concentrated for detecting cloud region. Firstly, Otsu's threshold algorithm is used to automatically separate the sky area from the others. Color index is conducted as the feature to tolerate the variation of light intensity in each day. In order to avoid the effect of diverse depths in the image, correction of the color index using virtual depth effect is required for correctly separating two classes in classification step. Supervised Bayesian Classification is performed to extract cloud area from the sky. Lastly, the cloud and sky areas are compared to calculate percentage of cloud amount in the sky. The experiments were conducted on rice field in Roi Et province, Thailand in November 2013. The result showed that our proposed method estimates cloud area effectively.