CN-115655144-B - Quantitative evaluation method for tea curl level

CN115655144BCN 115655144 BCN115655144 BCN 115655144BCN-115655144-B

Abstract

The invention discloses a quantitative evaluation method of tea leaf curl level, which comprises the following steps of S1, creating an objective evaluation method of the curl level of a measured tea sample to obtain an objective curl evaluation standard of the measured tea sample, S2, obtaining clear images of the front sides of shadowless tea leaves without overlapping each other, S3, utilizing python programming, adopting algorithms such as edge detection and skeleton detection in an OpenCV library to identify all target tea leaves, creating a measurement method of the curl degree of the tea leaves, S4, utilizing the objective curl evaluation standard of the measured tea sample established in the earlier stage to perform grading of curl such as spiral, curl, still curl and bending on each piece of tea leaves, and S5, obtaining the proportion of the curled tea leaves such as spiral, curl, still curl and bending in the tea sample and the average value of the curl degree. The invention defines the measurement method of the tea leaf curl degree, establishes an objective evaluation method of the tea leaf curl degree grade, and can realize digital evaluation of the tea curl degree so as to enable the evaluation to be more objective and accurate.

Inventors

LIN JIE
FENG HAIQIANG
LIANG XIUHUA
LI LAMEI
WANG WEIYI
WANG ZHOULI
ZHONG YU
TONG CHEN
MA JUNHUI

Assignees

浙江农林大学
丽水市茗天科技有限公司

Dates

Publication Date: 20260508
Application Date: 20220929

Claims (2)

1. A quantitative evaluation method for tea curl level is characterized by comprising the following steps: s1, creating an objective evaluation method of the curl level of the measured tea sample, and obtaining an objective evaluation standard of the curl level of the measured tea sample; s2, 50 tea leaves are collected at one time, so that the tea leaves are not overlapped with each other, vertical shooting is carried out through a shadowless lamp studio and camera equipment, and a clear image of the front surface of the shadowless tea leaves is obtained; S3, utilizing python programming, identifying all target tea leaves by adopting an edge detection and skeleton detection algorithm in an OpenCV library, and creating a measurement method of the tea leaf curling degree, and measuring the curling degree of the tea leaves according to the measurement method, wherein the calculation formula of the measurement method of the tea leaf curling degree is as follows: ; Wherein: CI is the curling degree; The total number of the pixel points of the tea leaf framework is calculated; the total number of the end points of the tea leaf framework is; SFD is the length of the framework diameter; s4, grading the curling degree of each piece of tea, such as spiral, curling, still curling and bending, by utilizing the curling degree objective evaluation standard obtained in the S1; S5, obtaining the proportion of curled tea such as spiral, curled tea and curved tea in the tea sample and the average value of the curling degree; in the crimping formula, the identification method of the skeleton end point adopts an eight-neighborhood discrimination algorithm, namely, all pixel points of the skeleton graph are traversed according to rows and columns from the first pixel point at the left upper corner of the graph, whether each pixel point is the skeleton end point is judged, if the pixel point is the skeleton end point, the skeleton end point is set to be P, reference pixel points in the neighborhood are point A and point B, whether the skeleton end point meets the following condition is further judged, namely, the sum of the pixel point numbers of the point A and the point B=1, dark square lattice represents skeleton points, light square lattice represents non-skeleton points, if the pixel point number is the skeleton end point number, the number is counted by 0.5, and the steps are repeated until the judgment of all the pixel points is completed; the skeleton length refers to the straight line distance of the two farthest points in the skeleton, the skeleton length represents the trend of a piece of tea leaf skeleton, if a certain skeleton point is consistent with the skeleton length trend, the tea leaf is straightened, and if a certain skeleton point is inconsistent with the skeleton length trend, the tea leaf is bent.
2. The quantitative evaluation method of the tea leaf curl level according to claim 1, wherein the skeleton is a central axis of an object, the skeleton is obtained, the main structure and shape information of the object are highlighted, redundant information is removed, and detection of feature points on an image can be achieved according to the information, wherein the feature points on the image comprise end points, crossing points and inflection points.

Description

Quantitative evaluation method for tea curl level Technical Field The invention relates to the technical field of tea quality control, in particular to a quantitative evaluation method for tea curl level. Background The curling degree of the tea is an important evaluation index of the appearance of curled tea, and the curling degree reflects the quality and the processing level of fresh tea to a great extent, for example, special grade Biluochun requires tea strips to be curled into snails, and special grade Mongolian sweet spot requires the appearance to be thin, elegant and even. The prior evaluation method has the technical defects that: 1. Mainly relies on manual sensory evaluation, requires professional personnel, and has no unified evaluation standard and evaluation method. 2. The evaluation is very subjective, not accurate and visual enough, and the digital evaluation of the crimping degree is not realized. Therefore, a digitized and objective evaluation system of the tea leaf curl degree is established, the tea leaf curl degree can be rapidly judged, the efficiency is improved, the resources are saved, the tea leaf quality is improved, and a reference basis is provided for digitized classification of tea leaves. Disclosure of Invention The invention aims to provide a digital evaluation method for the tea leaf curl, which can rapidly finish the digital evaluation work for the tea leaf curl, can be used for measuring the curl index of dry tea and also can be used for monitoring the curl index of products in real time in the tea processing process, and provides a reference for realizing digital processing so as to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: A quantitative evaluation method for tea curl level comprises the following steps: s1, creating an objective evaluation method of the curl level of the measured tea sample, and obtaining an objective evaluation standard of the curl level of the measured tea sample; S2, about 50 tea leaves are collected at one time, so that the tea leaves are not overlapped with each other, and a clear image of the front surface of the tea leaves without shadows is obtained by vertical shooting through a shadowless lamp studio and using shooting equipment; S3, utilizing python programming, adopting algorithms such as edge detection and skeleton detection in an OpenCV library to identify all target tea leaves, creating a measurement method of the tea leaf curling degree, and measuring the curling degree of 50 pieces of tea leaves (quantified by pixels) according to the measurement method, wherein the calculation formula of the measurement method of the tea leaf curling degree is as follows: Wherein: CI is the curling degree; The total number of the pixel points of the tea leaf framework is calculated; the total number of the end points of the tea leaf framework is; SFD is the length of the framework. S4, grading the curling of each piece of tea, such as spiral, curling and bending, by utilizing an objective evaluation standard of the curling degree of the measured tea sample established in the earlier stage; s5, obtaining the average value of the curling degree, such as the ratio of spiral, curling and bending, of the tea leaves in the tea sample. Further, the objective evaluation method for creating the curl level of the tea sample in the step S1 specifically includes the following steps: (1) Collecting a physical standard sample of certain types of tea, extracting a certain amount of samples according to requirements, and measuring the curling degree of the tea according to the method of the patent; (2) Judging whether the sample data obeys normal distribution; a. If the sample data is subjected to normal distribution, dividing the sample data into a plurality of grades according to the requirements according to probability theory and mathematical statistics principle, obtaining a grading threshold value, and combining a tea leaf curling degree big database (big data sampling, expert giving evaluation) constructed in the earlier stage to obtain an evaluation standard of the tea leaf curling degree; b. If the sample data does not obey normal distribution, the sample data is firstly converted into normal distribution or approximately obey normal distribution according to a certain method, a grading threshold is obtained according to the method, then the data is reversely converted to obtain an actual grading threshold of the sample, and an evaluation standard of the tea leaf curl is obtained by combining a tea leaf curl big database (big data sampling, expert giving evaluation) constructed in the earlier stage. Further, the level setting range of the normal distribution includes: if the array x approximately follows normal distribution x-N (μ, σ 2), then the set of values is generally considered to be normal distribution; an evaluation exceeding μ+σ as a, etc