農耕地景與野生動物棲地之現場調查是生態保育的重要工作，為確實瞭解農耕地景對農地生態之相關影響，往往需要大量人力到現場紀錄，耗費許多時間資源。透過無人機的高機動性可以進行大面積的農耕地景勘查，能夠快速並有效地建構地景生態資料，是永續經營管理農地生態的絕佳工具。本技術以無人機搭載高解析度光學鏡頭，結合智慧化影像辨識技術來分析與量化農耕地景類別。主要透由產製可供人工智慧訓練用的農耕地景影像，進行人工智慧多樣性地景細節特徵訓練，並將訓練好的神經網路模型協助農耕地景的分類。本技術適用於不同的地景類別，透過無人機的高機動性可以進行大面積的農耕地景勘查，降低非常多的現場調查時間與人力成本。

關鍵詞：無人機、農田生態、農耕地景、野生動物、人工智慧

On-site investigation of agricultural landscapes and wildlife habitats is an important task in ecological conservation. In order to truly understand the relative impact of agricultural landscapes on agricultural land ecology, it often requires a lot of manpower to record on-site, and consumes a lot of time and resources. Through the high mobility of drones, large-scale agricultural landscape surveys can be carried out, and landscape ecological data can be quickly and effectively constructed. It is an excellent tool for sustainable management of agricultural land ecology. This technology uses drones equipped with high-resolution optical lenses, combined with intelligent image recognition technology to analyze and quantify the types of agricultural landscapes. Mainly through the production of agricultural landscape images that can be used for artificial intelligence training, the artificial intelligence diversity landscape detail feature training is carried out, and the trained neural network model is used to assist in the classification of agricultural landscapes. This technology is suitable for different types of landscapes. Through the high mobility of UAV, large-scale agricultural landscape surveys can be carried out, reducing a lot of on-site survey time and labor costs.

Keywords: unmanned aerial vehicle, farmland ecology, farming landscape, wildlife, artificial intelligence