The Optical Characterization And Volume Measurement Of Plant Leaves

Plant leaves bear the heavy responsibility of photosynthesis and are also an important part of plants.In general,there are some differences in the morphology,structure and composition of different types of plant leaves,and the leaves of plants also change due to changes in conditions such as water,light and nutrients.In current agroforestry,we often judge the plant species and infer its current growth and development status by the shape and color of the leaves.With the development of optical detection technology,studying the optical characterization of various plant leaves can help us to quickly and accurately identify plants and determine the growth and development status,solving traditional methods requires a lot of manpower and material resources and poor timeliness,and provides a new platform for the development of modern agroforestry automation,precision,digitization,and informationization.In this paper,the combination of information optics and computer image technology is used to study the recognition of plant leaves under different growth conditions,leaf vein texture extraction of plant leaves under optical information processing and plant leaf volume measurement which based on binocular stereo vision.the main work of the paper is as follows:First,this paper optimizes the Alexnet convolutional neural network based on the current popular deep learning technology and the deep learning migration principle.We fully contact the actual situation in the application of life,and select the image data of plant leaves under different backgrounds and different growth and development conditions for deep learning,and the recognition rate after training can reach86.25%.Second,this paper is based on the idea that leaf vein texture will affect the recognition rate during leaf recognition.On the basis of optical information processing on the surface of plant leaves using the quadratic Fourier transform,we combine it with the Sobel operator edge extraction and the image value matrix,effectively extracting the vein texture of the front of cherry and ginkgo leaves,which can provide a reference for improving the recognition rate of plant leaves.Third,Plant volume measurement is an important part of forest carbon sink and greenhouse effect research.At present,plant leaf volume measurement is difficult and there are few related studies.This paper proposes a method for measuring plant leaf volume based on binocular stereo vision.First,we select two representative trees and measured the volume of the trunk portion.The error of the measurement results is within 5%.Then,based on this,the plant leaf mosaic model and the branch model are furthercombined,and the volume of the plant leaves is measured by mathematical modeling.The method has the characteristics of simple operation,small workload and fast calculation,and also enables remote measurement of plant leaf volume in a variety of different species and environments.