Soybean flower and pod falling phenotype investigation method

The invention provides a soybean flower and pod falling phenotype investigation method. The method comprises the following steps: controlling a robot to move to a target plant position; controlling the robot to collect to-be-recognized plant images of the target plant in the whole growth period; based on a completely trained target recognition model, performing recognition frame selection on flowers and pods in the to-be-recognized plant image, and determining category and position frame information of the flowers and pods in the to-be-recognized plant image; and based on the category and position frame information of the flowers and the pods in the to-be-identified plant image, counting the falling conditions of the flowers and the pods in the whole growth period of the target plant to obtain the flower and pod falling phenotype of the target plant. According to the invention, the efficiency and accuracy of flower and pod falling phenotype investigation are improved, and high-throughput phenotype data acquis