Utilizing Formosa II Imagery to interpret peanut and rice fields

• Main Authors: Pei-Chen Lai, 賴姵蓁
• Other Authors: Re-Yang Lee
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/29319097365312188212

碩士 === 逢甲大學 === 土地管理所 === 96 === In Taiwan, the utilization of the farmland is complex, especially for the period of the second rice crop cultivation. In order to earn more income, farmers often cultivate economic crops, such as peanuts instead of planting rice. The using of remote sensing farm classify research always focus on rice farm. Our investigation area is Tu-Ku and Ta-Pi in Yun-Lin county and using multi-temporal FORMOSAT-2 Satellite Images for recognition. We also add crops’ cultivated character and in-situ investigation to increase correct percentage. Traditonal classificaton use pixel based to classify. This method would cause classification mixed. In order to let the result to fit the actual situation using the multi-temporal images to regional based classify.The result in September is the best result for classifying peanuts and rice.Moreover to use NDVI image could recognize the none peanuts and rice zone. Besides, the research confirm the add regional texture conditons cuold increase the correct percentage. The classified result of total correct percentage is 85.19% and Kappa value is 0.78 in sample area.In empirical area’s result total correct percentage is 86.60% and Kappa value is 0.79.The total classified results have few difference and that prove this method’s usability.Finally we will supply the result to econmical,council of agriculture and research departments.