Study On Online Detection Method Of Chlorophyll In Rape Based On Computer Vision

Chlorophyll is an important physiological pigment of higher green plants,and its effect on green plants is self-evident.Chlorophyll is an indispensable basis in the photosynthesis process that green plants rely on.At present,there are many schemes for chlorophyll content determination,and many scholars are studying the determination of chlorophyll based on the direction of computer vision.However,most of the research based on computer vision requires higher requirements for visual equipment.At the same time,due to the fixed design software operating environment and other factors,there are some limitations in the research of chlorophyll content recognition based on computer vision.This paper mainly selects the mobile equipment as the visual terminal equipment of computer vision,and provides online real-time feedback,It provides a feasible solution for chlorophyll content measurement,data storage,display and visual analysis.This paper studies the color distribution characteristics of rape leaves in RGB,HSV and L*a*b*color space through image color histogram and K-means clustering analysis.Then,the image is processed by means of mean filtering and Gaussian drying,and the image is processed by morphological erosion.The mask of the image is obtained by color segmentation technology,and the target is segmented by image segmentation technology,Finally,through K-means clustering analysis,the feature variable of the leaf image is obtained.Compared with Photoshop software,the recognition rate of the feature variable reaches 97.82%.In this paper,through the study of rape leaves,using apple mobile phone to take pictures of rape leaves,get the color space parameters of the image,establish linear regression and neural network regression model.Python programming language combined with Open CV image processing software library was used to analyze the image information of rape leaves.After obtaining the image information of the blade,the MATLAB data analysis and simulation software and data analysis SPSS software package are used to analyze,process and clean the data.Then,the Python based machine learning tool scikit learn is used to train and test the model.The linear regression model with normalized color parameter r as input and SPAD value as output is established,and the neural network model with g,b,G/B+R,R/G and G-R as input and SPAD as output is established.Then,the output values of linear regression model and neural network model are taken as input Y₁ and Y₂ to establish a three-layer neural network,which contains two hidden layers and the number of hidden nodes is 2 and 1 respectively,L-BFGS is used to optimize the training model.The retrieval model of chlorophyll content in rape leaves based on computer vision was established.Finally,Python is used as the background language to establish the online detection system of rape leaf chlorophyll.The system has the functions of real-time,effective and data analysis,which provides a system solution for chlorophyll content measurement and data analysis.The main purpose of this paper is to design and establish an on-line detection and data analysis platform based on machine vision chlorophyll retrieval system of rape leaves.In order to solve the rigid requirements of equipment in the process of chlorophyll determination,the on-line real-time processing method is used for real-time feedback of field chlorophyll data and trend,so as to assist the production and scientific research of rape.