DR. MUTHONI MASINDE

Citation:

SM Mwagha MM. "APPLICATION OF COMPUTER VISION IN DETECTING SKY OBJECTS AS WEATHER LORE CONCEPTS ." Interim: Interdisciplinary Journal. 2016;15(1):1-17.

Abstract:

Weather lore is a body of informal folklore associated with weather prediction. Different types of weather lore exist and utilize almost all available human senses (feel, smell, sight and hear). Out of all the types of weather lore in existence, it is the visual or observed weather lore that is mostly used by indigenous communities to come up with weather predictions. Modern scientists also observe the sky to enhance their numerical weather prediction models. The visualization and representation of knowledge from sky objects (such as moon, clouds, stars and rainbow) in forecasting weather is a significant area of research. In order to realize the integration of visual weather lore knowledge in modern weather forecasting systems, there is a need to characterize and represent weather lore knowledge on visual sky objects. This paper reports on a method of approximating the presence of astronomical and meteorological objects in the sky. To achieve this objective, we designed detectors for the sky objects and score their presence and quantity in the sky panorama. The method of recognizing objects in images using image feature extraction techniques together with benchmark of similarity of extracted object against ideal objects (ground truths) was used. The results of this study reveal that our method is ideal in unlocking the extraction and computation of similarity of visual sky objects. The recommendation of this study is to use our method as a preprocessing task (using represented weather lore concepts) in the process of predicting weather outcomes and verifying visual based weather lore. Keywords: Representation; object similarity; object recognition; weather lore; bag of words; image features; classifiers

Full text link