請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86361
標題: | 應用影像及環境感測資料建立蘆筍嫩莖鮮重預測模型 Using Image and Environmental Sensor Data to Construct Models for Asparagus Spear Fresh Weight Prediction |
作者: | Whai-Ler Teng 滕懷樂 |
指導教授: | 劉力瑜(Li-Yu Liu) 劉力瑜(Li-Yu Liu | lyliu@ntu.edu.tw | ), |
關鍵字: | 蘆筍鮮重,線性回歸模型,非線性回歸模型,異速生長模型, spear fresh weight,linear regression model,nonlinear regression model,allometric model, |
出版年 : | 2022 |
學位: | 碩士 |
摘要: | 蘆筍為多年生宿根作物,屬高經濟價值作物。除了應用留母莖栽培方法,採溫室栽培的蘆筍相較於傳統戶外種植的蘆筍,較不受自然環境的影響且可以維持穩定產量。透過物聯網系統與多源數據的結合,我們希望以統計迴歸模型來預測溫室内蘆筍的重量,並應用於智慧生產過程和產量評估。 在這項研究中,我們根據從傳感器或影像中獲得的數據,透過線性回歸和非線性回歸建立的模型來預測蘆筍嫩莖的鮮重。研究中發現,以氣溫、空氣相對濕度和太陽日射量等環境因素為自變數,統計模式不能很好地預測蘆筍的重量,判定係數均低於 0.3。如果使用蘆筍嫩莖影像面積作為自變數,則嫩莖重量的預測得到改進,判定係數為0.3531,高於使用統計模型。總而言之,我們希望將傳感器數據與多種統計方法相結合,可以了解這些遙感(例如物聯網設備)對於農業智能生產的貢獻和潛力,同時也呈現出該研究受到的限制以及一些建議有助於改善預測的模型。 Asparagus is a perennial root crop, which has a high economic value. In addition to applying the cultivation method of leaving the mother stem, asparagus grown in the greenhouse is less affected by the natural environment and could maintain a stable yield compared with the traditional asparagus grown outdoors. Through the combination of Internet of Things (IoT) system and multi-source data, we hope to build models to predict the fresh weight of asparagus spear in Taiwan and applied it to smart production process and yield assessment. In this research, we used linear and non-linear regression models depending on data obtained from sensors or cameras to predict asparagus spear fresh weight. It was found that using environmental factors such as air temperature, air relative humidity and solar radiation as independent variables, the statistical models could not well predict the weights of spears with all R2 less than 0.3000. If using spear areas in the images as the independent variable, the prediction of the spear weights was improved (R2 = 0.3531). In conclusion, we hoped that combining the sensor data with multiple statistical methods, could give the readers a general and potential of these remote sensing contributed to smart production applied to agriculture, also showed some restrained during the research and some implications suggest. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86361 |
DOI: | 10.6342/NTU202202686 |
全文授權: | 同意授權(全球公開) |
電子全文公開日期: | 2022-09-02 |
顯示於系所單位: | 農藝學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-2308202210364500.pdf | 2.27 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。