Infrared Thermal Imaging and Statistical Analysis for Plant Stress Detection and Phenotyping

 Although California is the largest agricultural-producing state in America, frequent drought conditions severely impact the amount of product produced each year. High-value crops, such as avocados (Persea americana), are extremely sensitive to water deficits. These drought conditions are also the reason why many plants are endangered. Even though plant stress is a critical factor impacting a plant’s health, current plant stress measurements involve complicated measurement systems that are both destructive to the plant and inconvenient to use for farmers and plant preservation groups. There is an urgent need for a nondestructive and convenient technique for plant drought stress detection. Three series of experiments were designed and conducted to verify the practicality of using infrared thermal imaging technique to detect plant drought stress. The experiments were conducted on three different plants: Avocado (Persea americana), a drought sensitive plant; Shaw’s Agave (Agave shawii), an endangered plant; and the Toyon (Heteromeles arbutifolia), a drought tolerant plant, under different weather and irrigating conditions. The infrared thermal images were processed, and plant surface temperatures were statistically analyzed on three different plants, which demonstrated a strong correlation between surface temperatures and plant health condition and showed the feasibility of using infrared thermal imaging technique as an efficient and convenient tool for early plant drought stress detection. Based on this infrared thermal imaging technique, a model is currently being developed to automatically statistically analyze and accurately predict early drought stress in plants.

加利福尼亚州虽是美国第一大农业生产州，但频发的干旱灾情严重制约了其年度农产品产出总量。高价值作物如鳄梨（Persea americana）对水分亏缺极为敏感。干旱同样是诸多植物濒临灭绝的重要诱因。尽管植物胁迫（plant stress）是影响植株健康的关键因素，但当前的植物胁迫检测手段多依赖复杂的测量系统，不仅会对植株造成破坏，也不便用于农户与植物保护组织的实际作业。因此，业界亟需一种无损且便捷的植物干旱胁迫（drought stress）检测技术。
本研究设计并开展了三组对照实验，以验证红外热成像技术（infrared thermal imaging）用于植物干旱胁迫检测的实用性。实验选取了三种具有不同耐旱特性的植物：对干旱敏感的鳄梨（Persea americana）、濒危物种肖氏龙舌兰（Agave shawii），以及耐旱的冬青叶石楠（Heteromeles arbutifolia），并在不同天气与灌溉条件下完成测试。研究人员对采集到的红外热成像图像进行处理，并对三种植物的植株表面温度开展统计学分析，结果显示植株表面温度与健康状态之间存在显著相关性，证实了红外热成像技术可作为高效便捷的工具，用于早期植物干旱胁迫检测。
基于该红外热成像技术，目前正开发一款自动化模型，可自动完成统计学分析并精准预测植物早期干旱胁迫情况。