摘要： 本研究選定印度棗、番石榴等主要熱帶品牌水果進行品質分級指標建立研究,並研發水果內、外部品質檢測技術,進而開發國產線上型檢測精密分級機推廣應用為目標.研究利用機器視覺影像等檢測技術,建立水果外觀特徵實用品質分級指標,包含顏色、果型、表面特徵、大小、成熟度及蟲害、風害、鳥害等不良品偵測分級;並採用近紅外線偵測建立水果近紅外線反射光譜與其榶度、脆度等內部品質之關係.印度棗、番石榴水果品質分級指標建立及實用線上型精密分級機開發應用,可提昇產品的品牌形象,促進國產品牌水果分級品質的建立,提昇國產水果市場競爭力,並將分級資訊提供給農民,作為栽培管理改進之參考。This research tries to find the criteria and indices of Indian jujube and guavas for quality evaluation. These quality indices should include inside quality and appearance quality and should be applicable on sorting machine. Non-Invasive detecting technique i.e., machine vision will be applied to detect the appearance features, for instance, color, shape, surface characteristics, sizes, maturity, insect damage, climate damage and bird damage. The NIR (Near Infrared) technique will also be utilized to identify inside quality of fruits, such as sugar content and brittleness. By setting up grading criteria and developing sorting machine for Indian jujube and guavas, it can ensure the grading quality of fruits and upgrade the competitiveness in market. This grading information can be fed back to farmers to improve their skill on growing fruits.