人工智慧可加速達成聯合國永續發展目標（SDGs），例如：SDG 15的陸地生態（Life on land）利用物種識別和智慧物聯網的廣泛運用，追蹤陸地動物的遷徙、族群數量水準等活動，進而增強永續的陸地生態系統。據報導知，臺灣的鳥類占了全球鳥種的二十分之一，因此有許多賞鳥人士慕名而來，所以本研究將快速發展的人工智能應用於聲音辨識技術，先擷取鳥類聲音的樣本之特徵資料，並將其以AI深度學習中的卷積神經網路建立模組，將模組建置在APP期望能滿足眾多賞鳥愛好者的使用需求。為了探討臺灣特有種鳥類聲音辨識系統的服務需求，我們以有使用過APP應用程式經驗的年輕族群為測驗對象，引用服務體驗工程法為理論基礎，訪談與觀察探討使用者行為中的隱藏的意義，從歸納出臺灣特有種鳥類聲音辨識系統的服務需求。 根據服務體驗工程法中的五大構面進行訪談，並將訪談的資料匯整到五大模型中，分析在使用臺灣特有種鳥類聲音辨識系統的潛在的問題與需求。根據研究訪談的結果，發現使用者的需求： (1)目前辨識率為77%，需要再提升預測的正確率。 (2)系統設計上需要加強美工部分以及，資訊反饋的豐富度。 (3)改善聲音易容易被干擾的因素。以上三項可做為未來後續服務的主要依據。 Artificial intelligence can accelerate the achievement of the United Nations Sustainable Development Goals (SDGs), such as: SDG 15's Life on land uses species identification and the widespread use of the AIoT(Artificial Intelligence of Things) to track the migration of terrestrial animals, population levels and other activities, and then Enhance sustainable terrestrial ecosystems. According to reports, Taiwan's birds account for one-twentieth of the world's bird species, so many bird watchers come here. Therefore, this research applies the rapidly developing artificial intelligence to sound recognition technology. The characteristic data of the sample is used to build a module with the convolutional neural network in AI deep learning, and the module is installed in the APP to meet the needs of many bird watching enthusiasts. In order to explore the service needs of Taiwan's endemic bird sound recognition system, we took young people with experience in using APP applications as the test objects, cited the service experience engineering method as the theoretical basis, and discussed the hidden meaning in user behavior through interviews and observations. , summed up the service requirements of Taiwan's endemic bird sound recognition system. Interviews were conducted according to the five aspects of the service experience engineering method, and the interview data were compiled into five models to analyze the potential problems and needs of using the sound recognition system for endemic species of birds in Taiwan. Based on the results of the research interviews, the needs of users were identified: (1)The current recognition rate is 77%, and the accuracy of prediction needs to be improved. (2)The system design needs to strengthen the art part and the richness of information feedback. (3)Improve the factors that the sound is easily disturbed.The above three items can be used as the main basis for future follow-up services.
