Dangshan Pear's Fruit Stem And Size Consistency Recognition Based On Computer Vision Technology

Dangshan pear,one of the most popular specialties in China,has gained increasing attractions throughout the world.At present,artificial classification is the most frequently-used method in pear recognition.However,standards for pear recognition vary greatly,which directly lead to the insufficiency and error in recognition.In this study,computer version technique was applied to the automatic crisp pear recognition using pear as a case study.The research comprises: image acquisition and preprocessing,edge detection,size consistency judgments,and fruit stem region detection.Main issues and results obtained in this work include:Firstly,reprocessing methods of pear image was studied.According to the characteristics of pear,parameters for pear image reprocessing,such as grey level transformation,median filtering,morphological processing,was optimized.The optimal parameters were finally applied to automatic detection of pear characteristics.Then,based on the optimal size extraction of pear,mathematical model for size detection of pear image was established.An improved extraction method for maximum transverse diameter of pear was proposed after the comparison among commonly-used size detection methods for fruits.Judgment rules for pear size consistency was developed based on statistical features of pear transverse diameter.Additionally,the method was used to the extraction of pear transverse diameter,and the rule was applied to the consistency judgment of pear size.Finally,we first conducted an extraction of pear edge and centroid.Then the edge was equally divided.On the basis of pear fruit stem characteristics;we proposed a novel strategy for seeking fruit stem region through statistical edge division of pixel number.In this strategy,the outline was equally divided according to centroid point of bisection through the setting of reference point and direction in edges.After a statistical analysis of pixel information in each region,the accuracy in PEAR fruit stem extraction efficiency was enhanced.