Study On The Model Of Surface Characteristics And Water Content Change Based On The Surface Of Fresh Leaves Of Tea

Traditional detection method of water content includes drying and weighing method and sensor changed into electric quantity detection method.Although it is more accurate,it will damage the samples,and the detection rate is slow.The research of this paper is to realize the automation and intelligent tea green airing.Establishing the model of multi-spectral technique based on green tea green airing evaluation,and then the tea green airing process control will become accurater relatively.It has the important practical Significance to improve the quality of tea production and processing.The main research contents and conclusions are as follows:(1)Through analysis of the images of tea leaves,leaf image acquisition and pre processing method is put forward.Besides the tea geometric parameters,texture features and color space are discussed.And reasonable blade parameter measurement and texture features are proposed for value and color feature value algorithm,then obtain the image feature parameters and studying the relationship between these parameters and the tea water content.(2)Using digital camera to obtain tea leaf images.Analysis of how to obtain the target leaf using difference method,the noise source of image and image of two value process and expansion process,the removal of acnode noise by median filtering.In the stage preprocessing,according to the features of leaf images obtained,getting the two value images by threshold segmentation method.And then using the two value morphology methods to improve the image quality on the basis of this technology,Through the test of the various detection algorithms were compared,the experiment proved that the Canny operator has more advantages.On this basis,the blade geometry parameters,the texture and color characteristic value analysis are realized.(3)The process of tea spreading test and the changes of fresh activity of leaf in tea leaves were studied and how to realize the change of the water content and the fresh degree of real-time monitoring and analysis of the fresh tea leaves in tedding process.Using the digital camera and tea moisture tester and other test equipment to collect 30 samples of 540 water cut.Through the changes of various characteristic parameters and the color of tea leaves to study the change of water content and the process of tedding fresh degree.Established the relationship model of blade 13 parameters and moisture content,it is concluded that leaves area,perimeter,leaf width,S value,consistency,and entropy of seven parameters is superior to the correlation of R,G,B,H,V,and the mean.Then the related amount and exclude the unrelated parameters.(4)Moisture content was established based on principal component analysis,and between the two big main component models,and established the moisture content of fresh tea leaves and comprehensive area,perimeter,leaf width,S value,consistency,and entropy of seven parameters of the standardization of data forecast model.Using the return of the moisture content in the green process value compared with the test results,and the error analysis,validate the accuracy of the model and the accuracy is above 90%.