北美對於娛樂大麻與醫療大麻的推動，促使高經濟作物的種植盛行，進而帶動LED植物燈以及相關種植設備的發展與導入，加上極端天氣變化頻繁以及疫情因素，更推升高經濟作物的市場需求。本論文針對高經濟價值之食用植物為標的，希望結合感測器與AI演算法開發出智慧型非破壞手持式光合色素分析儀，以建立LED照射光譜與光合色素的關聯性。所開發出的感測裝置可應用於預估植物生長情況，提供植物工廠栽種者於種植時的參考。 本研究使用番茄、萵苣、羽衣甘藍及皇宮菜為樣本。以HPLC量測結果作為基準，比對由手持式光合色素分析儀感測得到的葉綠素a、葉綠素b、β胡蘿蔔素、葉黃素四個光合色素之比例。感測所得的訊號，透過訊號特徵提取與AI模型的建立，提高非破壞手持式光合色素分析儀的精準度，根據實驗結果顯示，本儀器所量測到的各個經濟作物之數據與HPLC所量測到的結果差異皆在10%以內，證明以光合色素分析的方法測定植物營養素的準確率高達90%以上，符合原先之設計目標。 ;The promotion of recreational and medical marijuana in North America has spurred the widespread cμLtivation of high-value crops, driving the development and adoption of LED grow lights and related equipment. Additionally, frequent extreme weather and the pandemic have further increased market demand for these crops. This paper focuses on developing a smart, non-destructive, handheld chlorophyll analyzer using sensors and AI algorithms to establish the correlation between LED light spectra and chlorophyll content. The device can estimate plant growth conditions, providing usefμL data for plant factory growers. Tomatoes, lettuce, kale, and choy sum were used as samples. HPLC measurements served as the benchmark, against which the proportions of chlorophyll a, chlorophyll b, beta-carotenoids, and xanthophylls measured by the handheld analyzer were compared. By extracting signal features and building AI models, the accuracy of the analyzer was improved. Experimental resμLts show that the data from the analyzer differed from HPLC resμLts by less than 10%, achieving an accuracy rate of over 90%, which meets the initial design goals.
