Bruise detection of apples based on passive thermal imaging technology

Abstract

It is crucial for early detection of mechanical damage to improve the quality of commodity fruits. To address the issues related to the fuzzy damage characteristics of early fruit damage and the challenges associated with accurately identifying the damaged locations, this study proposed a method of detecting the initial damage of apples based on passive thermal imaging technology. Passive thermal imaging technology was used to collect the thermal images of the apples 0–60 min after damage, once every 2 min, for a total of 5 times. The temperature difference data of the damaged part and the undamaged part were extracted from the temperature data, and the evolution law of the temperature difference with time was analyzed. Then, it was transformed into the area under the curve distribution data of the area under the curve, so as to generate the corresponding gray image for realizing the detection of the initial damage of apples. The results showed that the detection precision of the undamaged apples was 98.3%, and the average detection precision of the damaged apples was 86.7%. The proposed detection method can realize the nondestructive detection of the initial damage of apples, and provides research ideas and theoretical basis for more reliable detection of the initial minor damage of apples in the future.