Netted melon fruit phenotype extraction and quantization method

Abstract

The invention discloses a netted melon fruit phenotype extraction and quantization method. The netted melon fruit phenotype extraction and quantization method comprises steps as follows: step 1, according to a reference substance method, a red green blue (RGB) image and a depth image of netted melon fruit is collected through a RGB-depth (D) visual sensor, and the RGB image is preprocessed; step 2, the preprocessed image is divided, a reference substance image is obtained, and the pixel length ratio is computed; step 3, the fruit is divided based on the depth image, and shape feature parameters of the fruit is obtained by computing with the pixel length ratio; step 4, outline information of the fruit obtained from the depth image is used in the RGB image to obtain a RGB image area of the fruit; step 5, according to the RGB image area of the fruit, color space switching is utilized to obtain color parameters H and S, and net characteristics of a melon is quantized based on the RGB image area and taking advantage of the net covering rate, a gray level co-occurrence matrix parameter and a polychotomy dimension parameter. By means of the netted melon fruit phenotype extraction and quantization method, extraction comprehensiveness and obtaining efficiency of netted melon phenotype are improved, and objectivity and accuracy of netted melon phenotype quantization are improved.

Description

technical field [0001] The invention relates to an objective quantification method of reticulated melon fruit phenotype, in particular to a method for extracting and quantifying the phenotype of reticulated melon fruit throughout the growth period based on RGB-D visual sensors, and in particular to a method for extracting and quantifying the phenotype of reticulated melon . Background technique [0002] Reticulated melon phenotype is an important indicator for breeding and identification of new varieties. At present, the phenotype of reticulated melon is mainly obtained by manual visual inspection and manual measurement to characterize the phenotype of reticulated melon. Visually inspect the color of the skin (whether the skin turns yellow, whether the fruit is cracked, whether the skin is double-colored), the distribution of reticulation (sparseness), whether there is wrinkled skin, and use a vernier caliper to measure the size of the vertical and horizontal diameters, etc...

AI Technical Summary

Problems solved by technology

It is impossible to extract all the phenotypic parameters of reticulate melon during the growth and development process under the natural growth state in the algorithm

Images