A research on determining sunflower planted area by remote sensing technique in different planting zones

The use of remote sensing techniques to agricultural purposes is rapidly transferred to the applications in the world and satellite images used for this purpose. Remote sensing techniques have remained mostly consists of trial in our country. Sun flower plants, which are one of the essential oil plants having important economic and nutritional value among the agricultural products in the world, are grown up in large areas of our country. With these characteristic, sunflower has a very important place in national and international agricultural policy. In order to establish policies for incentives, direct income support to agricultural production, etc., accurate information is needed about cultivated areas. Monitoring oil plant sunflower production, having national and international importance, will meet an important missing in our agricultural statistics. Improving the accuracy rate of determining sunflower planted areas and yield harvest, which is currently identified by traditional methods and declaration of farmers, is the main purpose of this project. For this purpose, the most suitable method is seen as remote sensing techniques. With this study all the parameters for sunflower crop determination from satellite images will be examined. In our county, sunflower plants are grown in two different planting times as the first or second crop. While sunflower cultivation is preferred as the first product in the Thrace region, it is preferred as the second crop (after wheat) especially in the large areas of the Aegean region. This case causes to gather images at different times at the level of regions as well as the varieties of crop types that may be mixed with sunflower plants. With this study, to determine sunflower planted areas, some parameters such as growth stage of vegetation, densely coverage time of vegetation, effects of vegetation and soil reflection, possible mixing vegetation such as corn, etc., with sunflower plants on the level of regions will be defined. Depending on this information the most appropriate classification method will be determined for cultivated area. In this research, SPOT 5 and Sentinel 2A satellite images were used. Classification methods such as pixel-based (band ratio, Para ML and NDVI) and object-based (segmantation) classification methods was applied on the images and as a result eventually it was determined that the most appropriate classification method of supervised classification, Para - ML.