Objective:To investigate the relationship between the development and traits of Forsythia fruit,and establish a method for identifying the quality of Forsythia fruit based on computer vision technology.Methods:Python library LabelMe was used to annotate the micro-cross-sectional images of Forsythia fruit,and the area proportion of each part of the micro-cross-section of Forsythia fruit was obtained.The self-made software Lq_measure was used to measure the length,width,and single fruit area of Forsythia fruit.At the same time,the quality of Forsythia fruit peel(including septum)and seeds was measured,and a dataset was created.Results:There were significant differences in the traits of Forsythia fruit,which confirmed the feasibility of using computer vision technology to screen high-quality Forsythia fruit and had prospects for industrial application.A set of experimental operation procedures for the study of Forsythia fruit traits based on Lq_measure software was established,provid-ing a basic method for the establishment of a Forsythia fruit trait database and reliable technical support for the theoretical research of"discriminating shape and discussing quality"of traditional Chinese medi-cine based on Forsythia fruit shape.Conclusion:The trait of Forsythia seed heart could be used to iden-tify the fruit maturity of Forsythia.The correlation between the length-width ratio coefficient of For-sythia and the proportion of Forsythia fruit peel weight was 0.74,providing a theoretical basis for fu-ture Forsythia quality control.
NETL
NSTL
万方数据
