應用氣象雷達於生物影像辨識技術之開發-以台灣過境猛禽為例

Using Doppler Weather Radar to Recognizing Raptor Signal during Migration Seasons in Taiwan

王佑仁(Wang Yu-Jen)

指導教授 ： 賴信志

長榮大學/資訊暨工程學院/資訊管理學系(所)/碩士(2012年)

https://doi.org/10.6833/CJCU.2012.00030

摘要

台灣受到春、秋兩季季風影響，造就許多生態特殊景觀。赤腹鷹與灰面鵟鷹的遷徙，是在世界上相當特殊的奇景；也使得台灣成為世界上研究猛禽生態的少有舞台。本研究主要目標是應用氣象雷達偵測猛禽遷徙過境訊號。從氣象雷達所產出的雷達回波圖訊號進行即時性分析，並判別雷達圖中的訊號是否為猛禽遷徙時所留下的訊號。
本研究發展之辨識方法是使用去除訊號的方式來進行辨識。辨識系統會將雷達圖中的雲層訊號與水氣訊號去除，當所有水氣訊號及雜訊都被去除後，再從剩餘訊號中判別猛禽訊號。為配合猛禽遷徙之飛行特性，本研究使用雷達掃描仰角為〖0.5〗^° 角與〖1.4〗^° 角，分別對這兩種雷達仰角的訊號，進行訊號辨識並找出猛禽訊號。
本研究使用的雷達資料從2009年04月到2011年11月為止總共辨識了994,674張雷達圖，並從多達90萬張雷達圖中，找出數量相當稀少的猛禽過境雷達訊號個案。從這些猛禽過境訊號個案中可以了解，應用本研究所發展之辨識方法，進行猛禽訊號辨識是可行的。本研究也是自有文獻紀錄以來，首次開發出自動辨識猛禽訊號之研究，從三年雷達資料辨識結果中，可得到單季最佳辨識率為86.88%。本研究也從辨識結果中整理出四天猛禽大量遷徙時，從氣象雷達圖中所辨識出的猛禽訊號遷徙個案，計算飛行速度與分析遷徙路徑。本研究所發展的應用氣象雷達圖辨識猛禽訊號之方式，將有助於初步預測灰面鵟鷹與赤腹鷹之遷徙路徑，以達到保育這兩種東亞特有遷徙性猛禽之目的。

關鍵字

氣象雷達 ； 猛禽 ； 生物影像辨識 ； 猛禽遷徙資訊系統

並列摘要

Spring and autumn monsoon affect Taiwan so much that they create many special ecological landscapes. In these landscapes, the migration of Chinese Goshawk and Grey-faced Buzzard is a very exceptional spectacle, and this also enables Taiwan to become a rare stage to study the ecology of migratory raptors in this world. The main object of this study is developing the application technique of Doppler weather radar to detect the signals of migratory raptors. We will analyze immediately the radar echoes signals which output from the weather radar and determine whether these signals in the radar diagram are the signals which are left in the migration of the birds or not.
The identification method, Raptor Identification System, developed in this study is using the signal removing method to identify the migratory raptors. This system will remove the cloud and water vapor signals in the radar diagram, and it will distinguish signals of migrating raptors from the remaining signals after removing all traditional water vapor signal and noise. In order to tie in with the flight characteristics (200m ~ 1200m) of migratory raptors, this identification system uses the weather radar to scan in the elevation angle of 0.5 degree and 1.4 degree to detect the signals of migratory raptors.
This “Raptor Identification System” has identified a total number of 994,674 weather radar diagram from April 2009 to November 2011. The few cases from the large number raptor migration signals have been identified from nine hundred thousand radar diagrams. These successful identification cases have showed that it’s feasible to apply Doppler weather radar to detect raptor migration signal during the migration season. The automatic “Raptor Identification System” is first development of all recorded time. Successful identification rate is 86.88% in a season statistics for three years. There are four cases that it’s the large number migration signals identified from radar diagrams and calculate raptors fly speed and fly way. The “Raptor Identification System” will help to understand the mystery of fly way for Chinese Goshawk and Gray-Faced Buzzard, and it also can help for the enforcement of conservation to achieve the purpose of conservation during the migratory sessions.

並列關鍵字

weather radar ； raptor ； biological image recognition ； raptor migration information system

參考文獻

[12]鄭育昇、孫元勳、鄧財文(2006)，利用氣象雷達探討2005年秋季赤腹鷹過境恆春半島之模式，台灣林業科學 21:4，491-198頁。

[4]陳韻如、孫元勳、鄧財文(2007)，2005年春季灰面鵟鷹在台中南部之北返遷徙模式，國立屏東科技大學，碩士論文。

[1]Bamford, M, D. Watkins, W. Bancroft, G. Tischler and J. Wahl. (2008), Migratory Shorebirds of the East Asian - Australasian Flyway; Population Estimates and Internationally Important Sites, Wetlands International Oceania. Canberra, Australia, pp.158~pp.199.

[2]Eastwood, E.(1967), Radar Ornithology, Methuen, pp.278.

[4]Gauthreaux, S. A., Jr.(1970), Weather radar quantification of bird migration, Bioscience 20, pp.17-20.