Design And Implementation Of Image Software Based On Mobile Terminal For Parameters Measurement Of Plant Leaf

As an important part of plants,the shape parameters of leaf blades are important research contents of botany.Leaf shape parameter is not only one of the important basis for distinguishing plant species,but also an important index to measure plant growth status,predict production situation,manage crop cultivation,prevent and control pests and diseases.Therefore,it is very important to accurately measure the shape parameters of the blades for the study of plant growth and the choice of environment for the cultivation of fine varieties.Traditional methods of measuring leaf geometrical parameters,such as Grid method,copy weighing method,low efficiency,poor accuracy,the leaf area instrument method is expensive,not easy to carry and maintenance in the field,and can only measure the leaf area,the image processing scanning method has high precision and low cost,but the equipment is complex,it can't timely obtain the measurement result in the field.Aiming at the disadvantages of these methods and combining the advantages of software and hardware of smart phone,this paper designs and realizes the measurement software of leaf shape geometry parameters based on image processing technology on Android smart phone platform,which improves the efficiency of data acquisition,easy to carry,convenient to operate,high measurement precision and real-time.This thesis mainly studies the measuring methods of the parameters such as perimeter,area and middle axis of leaves.The image acquisition,gray processing,threshold processing,separation of leaf and reference,geometric correction of leaf and the removal of leaf petiole are briefly described,and the measuring method of the perimeter and area of the leaf is simply described,and the measuring method of the vane axis is emphasized.According to the two key steps of skeleton extension and search centre corridor diameter in leaf shaft detection,the skeleton expansion method based on distance transformation and the Fast Search Centre Corridor path method are proposed.The skeleton expansion method based on distance transformation is to expand each skeleton endpoint that is not falling on the leaf boundary to its nearest boundary point.The improved method of exhaustive skeleton path is to establish skeleton tree structure and to find the path of each endpoint to the axis starting point recursively.The fast search centre corridor method firstly removes most of the non-middle axis skeleton branches in the leaf skeleton by the method of removing non-middle axis skeleton branch based on the perimeter dividing criterion;Then,the improved exhaustive skeleton path method is used to obtain all skeleton paths,and the candidate centre corridor path is searched according to the area dividing criterion;Finally,the centre corridor diameter is determined based on the maximum curvature criterion.Among them,the method of removing non-middle axis skeleton branch based on the perimeter criterion is to calculate the boundary length ratio of the skeleton endpoint to the beginning of the middle axis clockwise and counterclockwise,then remove the skeleton branch of the endpoint,which can quickly reduce the number of leaf skeleton path.The improved method of exhaustive skeleton path is to establish skeleton tree structure and to find the path of each endpoint to the axis starting point recursively.Compared with the standard method,the experimental results show that the framework expansion method based on distance transformation and the fast search centre corridor path method not only guarantee high accuracy of measurement,but also improve the detection speed of the middle axis significantly.At the end of this paper,the development environment,function frame,interface layout and operation flow of the software are given,and the operation methods and implementation modes of the software are specified.