Intensive seedling cultivation technology is a crucial step in vegetable production,the quality of seedlings directly affectsthe growth and cropyield in the later growth stages.Seedling quality varies due to different production techniques,how toevaluate the quality of a seedling has become a concern for producers and users.Traditional methods of evaluating seedlings mostlydepend on subjective judgments based on appearance,lacking objective and accurate quantitative evaluation criteria.This study utilizes a plant phenotyping measurement system to capture multi-angle two-dimensional image sequences of celery,lettuce,and tomato seedlings.Based on the acquired three-dimensional point cloud models of the plants,phenotype data for vegetable seedlings are extracted.A model between phenotype data and seedling vigor indicators is established using a random forest regression algorithm,facilitating the reconstruction of the three-dimensional structure of individual vegetable seedlings.The results indicate that the lettuce model exhibits a strong correlation with an R2 value of 0.91,while the celery and tomato models have R2 values of 0.83 and 0.77,respectively.The tomato model shows the lowest rRMSE is 0.33,indicating that 3D structural modeling based on vegetable seedlings can provide scientific basis for vegetable seedling evaluation system.
维普
万方数据