查詢結果分析
來源資料
相關文獻
- 結合SSM/I衛星資料與類神經網路推估海面上颱風降雨量之研究
- 結合群聚分析與人工智慧於颱風時雨量預測
- 應用類神經網路模式模擬颱風降雨量
- Neural Network Procedures for Taguchi's Dynamic Problems
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
- 專家系統振動訊號圖型判別之研究
- 反傳遞模糊類神經網路於流量推估之應用
- 類神經網路(Neural Networks)的種類及其在影像處理上的應用
- C++Fuzzy類神經網路物件導向發展系統之建立
- 臺灣汽保費率之估計--對數線性費率模式與類神經網路之比較
頁籤選單縮合
題 名 | 結合SSM/I衛星資料與類神經網路推估海面上颱風降雨量之研究=Study on Estimating Typhoon Rainfall over Sea from SSM/I Satellite Data Utilizing Artificial Neural Networks |
---|---|
作 者 | 魏曉萍; 葉克家; 劉振榮; 趙俊傑; | 書刊名 | 大氣科學 |
卷 期 | 36:2 2008.06[民97.06] |
頁 次 | 頁147-162 |
分類號 | 328.888 |
關鍵詞 | SSM/I衛星資料; 類神經網路; 颱風降雨量; SSM/I satellite data; Artificial neural network; Typhoon rainfall; |
語 文 | 中文(Chinese) |
中文摘要 | 每年颱風所帶來的豪大雨常帶給台灣地區民眾生命及財產損失,尤其一些停滯在台灣地區較久颱風,如納莉及桃芝等皆造成大規模之淹水與土石流災害。因此定量降雨預測是目前防災上重要課題之一,而氣象衛星資料估算降雨在此課題上扮演著重要之角色。本研究即使用氣象衛星上之SSM/I儀器估算海面之降水量。早期推估降雨量最常使用方法有物理法與統計法兩種。物理法之模式常加以簡化,以致無法完全符合實際大氣狀況;統計法雖然具有簡便迅速優點,但常依特定季節與地區建立亮度溫度與降雨量間之關係。因物理法與統計法有上述之缺點,遂有利用人工智慧相關理論而為另一種替代方法。本文採用四種類神經網路,及三種不同SSM/I微波頻道之組合,嘗試建立適用於海面上推估颱風降雨量之方法,並以綠島、蘭嶼、彭佳嶼及東吉嶼等4個觀測站資料進行驗證。研究結果顯示,所建立之類神經網路以利用第三種SSM/I微波頻道組合,可較準確預測海面上之颱風降雨量;在四種類神經網路中,以FNN及AASN方法較適合於海面上颱風降雨量之推估。 |
英文摘要 | Heavy rainfalls resulted from the typhoons cost human lives and financial damages in Taiwan every year. This is especially the case for stagnant typhoons over Taiwan area such as Nari and Toraji, which bring large scale disasters in forms of floods and debris flows. Therefore, one of the most important topics in disaster prevention in Taiwan would be the prediction of precipitation, and rainfall estimation using meteorological satellite data plays an important role in this topic. In earlier times, physical and statistical methods are the two most frequently used methods for estimating rainfall. The physical method is usually over-simplified and its results are incoherent with the realistic atmospheric conditions, while the statistical method often over-relies on the relationship built according to a specific season and region, despite its advantages of simple and fast computation. Since the physical and statistical methods have the limitations described above, theories related to the artificial intelligence arise to be an alternative method in recent years. This research utilizes four types of artifical neural networks, combining with three different combinations of SSM/I microwave frequency channels, in an attempt to establish a suitable method for estimating rainfall at sea surface during typhoon periods. These results are then verified with the data from four rainfall stations located at Lyudao, Lanyu (Orchid Island), Peng Chiayu, and Tungchityu. From the results, it is found that data from the third type of combination of SSM/I microwave frequency channels is more accurate at estimating the rainfall in typhoon periods. Among the four types of artifical neural networks, the FNN and AASN methods are more suitable for estimation of sea-surface rainfall in typhoon periods. |
本系統中英文摘要資訊取自各篇刊載內容。