Method for identifying cordyceps sinensis in complex environment

The invention discloses a method for identifying cordyceps sinensis in a complex environment. The method specifically comprises the following steps: step 1, acquiring target image information; 2, preprocessing the target image; step 3, optimizing the YOLO v3 deep neural network model; 4, identifying a cordyceps sinensis target in the pre-target image, and extracting an ROI (Region of Interest) region of the target; step 5, extracting a minimum enclosing rectangle of a cordyceps sinensis target in the ROI according to the inherent attribute and the feature information of the cordyceps sinensis target; and the specific position information of the cordyceps sinensis target is obtained. According to the identification method, the identification efficiency of the special plant can be improved to a certain extent, manual labor needing to be intervened is reduced, and certain practical significance is achieved. 本发明公开了复杂环境下冬虫夏草的识别方法，具体按照如下步骤进行：步骤1：采集目标图像信息；步骤2：对目标图像做预处理；步骤3：优化YOLO v3深度神经网络模型；步骤4：识别预目标图像中的冬虫夏草目标，并提取该目