Plant leaf segmentation method and system using synthetic data

Abstract

The invention discloses a plant leaf segmentation method and system using synthetic data. The method comprises the following steps: constructing a plurality of blade three-dimensional models of different postures and different colors of the blade image based on the blade image; projecting the blade three-dimensional model to a two-dimensional plane to generate blade two-dimensional image data, andfusing the blade two-dimensional image data with different background images to obtain a training set; training the deep learning model through the training set to obtain a blade segmentation model;and inputting the to-be-segmented image containing the blade into the blade segmentation model to obtain a segmented blade image. The three-dimensional leaf model is used for generating segmentation data; the method comprises the steps of obtaining a plant leaf image; the leaf models of different shapes and different colors are obtained through reconstruction and are fused with different backgrounds, a large number of training pictures and training labels are automatically generated to form a training set, the energy consumption of manual image marking is reduced, plant leaf images can be fully automatically segmented under the natural background condition, and the segmentation effect is good.

Description

technical field [0001] The invention relates to the field of image segmentation, in particular to a plant leaf segmentation method and system using synthetic data. Background technique [0002] Studying plants, the identification, detection and segmentation of plant leaves is an important task for us in computer vision of plants. However, the background conditions of plant leaves under natural conditions are generally complex, and the types of plant leaves themselves have lines, colors, and textures. Therefore, it is a very difficult task to identify, detect, and segment plant leaves. Machines can effectively identify leaves, based on the ability to extract plant leaves from the natural background. Subsequent analysis of the plants can only proceed if the leaves are extracted from the background. For example, the segmented leaves can be used to determine whether the plant is growing healthily and whether it has suffered from diseases and insect pests. The segmentation alg...

AI Technical Summary

Problems solved by technology

However, based on the method of deep learning, there must be training data. At present, the public data sets in the world are manually labeled. Manually labeled data means that it consumes a lot of human and financial resources, and manual labeling is also easy. labeling error

Therefore, the lack of data has seriously hindered the progress of deep learning methods in image segmentation algorithms.

And so far, there is no segmentation dataset specifically for plant leaves

At the same time, the lack of data also hinders the development of deep learning

[0005] It can be seen from the above that using the traditional leaf segmentation method requires manual input of some parameters, and automatic segmentation cannot be completed, or the segmentation effect is poor.

The segmentation method using deep learning can achieve fully automatic segmentation of leaves, but requires a large amount of training data

Images