Identification Of External Quality Grading Of Golden Dates With Multi-colour Model Fusion

Jujubes are a specialty fruit in China,and the golden-silk jujube in its family has rich nutritional value and unique medicinal value,but due to its crude grading and screening techniques,graded products are of mixed quality and difficult to get out of the international market.Using the local standard for golden-silk jujube as a guide,we visited farmers,jujube companies and experts in the field on the basis of field experiments to expand some of the attribute characteristic indicators of golden-silk jujube and propose more specific grading standards for colour,size and shape attributes.The research object was selected to collect images of golden jujube from multiple regions,varieties and environments,analyse,apply and optimise advanced algorithms and techniques,and build a grading system for the external quality of golden jujube based on the extended grading criteria,based on image processing,defect recognition and feature grading with the fusion of multiple colour models.The process research focuses on the following key technical challenges.Firstly,the use of multi-colour models for the external inspection of golden jujube fruit was investigated.Based on GB 5835-86,the common defect types of golden jujube are identified,the conversion of jujube images from RGB colour model to HSV colour model is processed,the defective jujube is identified based on the difference in the grey scale variation range of the image regions,and the defective location is precisely located using the region growth algorithm,and then the defective part is segmented out from the image by the Canny algorithm,and then the defect detection is completed.The second is a study on the grading of the external surface features of golden dates based on a multi-colour model.After pre-processing the date images,the RGB colour model was used to analyse the proportion of the colouring area to the fruit area as the colour grading criterion,and the Lab colour model plus the four-line scanning method was used to measure the size of the golden-silk dates to construct a block improvement algorithm for overall grading according to a certain threshold.The experimental comparison results showed that the block improvement algorithm improved the judgement and recognition efficiency of golden silk jujube compared with the original algorithm,and the judgement and recognition time of special grade fruit,first grade fruit,second grade fruit and third grade fruit was reduced by 0.0563 s,0.0208,0.0105 s and 0.0013 s respectively.