کاربرد پردازش تصویر میکروسکوپی در تشخیص و طبقه‌بندی اسپورهای تریکودرما

Trichoderma is a type of fungus that is commonly used in agriculture to prevent and control various plant diseases. One of the mechanisms that Trichoderma helps control plant diseases is by producing antibiotics that prevent the growth of pathogens. The ability to identify Trichoderma spores is essential to detect the presence of this fungus in plants and soil. However, traditional microscopic techniques used for spore detection are time-consuming and require the intervention of skilled technicians, leading to delayed detection and ineffective control measures. Methods based on microscopic image processing have been developed to smarten up the spore detection process, leading to faster and more accurate results. Although at present, researchers have widely used image processing to introduce intelligent methods in various stages of agriculture and horticulture, including planting, weed control, irrigation, spraying, fertilizing, monitoring plant growth and harvesting, using microscopic image processing can serve as a new tool in plant pathology. In this study, a new method for identifying and classifying the spores of three Trichoderma species including T. harzianum, T. atroviridea, and T. virens was developed using microscopic image processing techniques. For this purpose, a total of 100 microscopic images of these species were prepared and then the genetic algorithm was used to identify the most effective visual features for the classification of spores, including color, texture, and shape. The genetic algorithm is a meta-heuristic optimization algorithm that, based on an iterative process, identifies the most suitable inputs of machine learning algorithms to either minimize or maximize the output of the machine. In this research, the input and output of this algorithm were the extracted features and the prediction accuracy of the machine, which had to be maximized. After finding the effective features, the random forest was used to classify the spores, and an approximate accuracy of 95% was obtained. Furthermore, this study showed that image texture is the most effective visual feature for classifying spores. The use of such new methods can help control plant diseases more effectively by providing faster and more accurate detection of Trichoderma spores in plants and soil.