CN-121997202-A - Construction method and application of sea cucumber raw material hardness grading model based on machine learning

CN121997202ACN 121997202 ACN121997202 ACN 121997202ACN-121997202-A

Abstract

The invention discloses a construction method and application of a sea cucumber raw material hardness grading model based on machine learning, and belongs to the technical field of food processing. The construction of the sea cucumber raw material hardness grading model comprises the steps of obtaining sea cucumber raw material hardness evaluation index data, taking the sea cucumber raw material hardness evaluation index data as a training data set, taking one or more of sea cucumber moisture content, sea cucumber protein content, average reflectance value of sea cucumber at 960nm characteristic wavelengths and sea cucumber glycosaminoglycan leaching amount as training model input data, taking a corresponding hardness grade label as a training model training target, training a machine learning algorithm to obtain a sea cucumber raw material hardness grading model, wherein the grading accuracy of 16 prediction sets of experimental samples is 100%, and in subsequent application, the main reflectance interval accuracy of the grading model at 960nm of sea cucumber is 98.2%, and the model is more general in extreme cases.

Inventors

DONG XIUPING
Cai Houde
JIAO JIAN
ZHANG XU
WU QIONG
WANG ZHEMING
WANG HUIHUI
LI XINCAN

Assignees

大连工业大学

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (10)

1. The method for constructing the sea cucumber raw material hardness grading model is characterized by comprising the following steps of: acquiring evaluation index data of the hardness of the sea cucumber raw material, wherein the evaluation index data comprise the water content of the sea cucumber, a hardness grade label, an average reflectance value of the sea cucumber at 960 nm characteristic wavelength, the protein content of the sea cucumber and the dissolution amount of the sea cucumber glycosaminoglycan; taking the sea cucumber raw material hardness evaluation index data as a training data set; One or more evaluation index data of sea cucumber moisture content, sea cucumber protein content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and sea cucumber glycosaminoglycan leaching amount are used as training model input data, corresponding hardness grade labels are used as training model training targets, and a machine learning algorithm is trained to obtain a sea cucumber raw material hardness grading model.
2. The method of claim 1, wherein the machine learning algorithm is a random forest classification method.
3. The method of claim 1, wherein the random forest classification method uses Scikit-learn library of Python to construct a random forest classifier, sets the number of decision trees to 100, and defaults the rest of parameters.
4. The method according to claim 1, wherein the training model input data is sea cucumber moisture content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and/or sea cucumber glycosaminoglycan elution.
5. The method according to claim 1, wherein the hardness grade label is constructed by inviting experienced sea cucumber processing workers to perform sensory evaluation on the hardness of the sea cucumbers in an environment with uniform light and constant temperature through independent touch and pressing modes, grading according to four predetermined grades of standards, re-evaluating until consensus is achieved when expert opinion is inconsistent, and finally assigning a recognized sensory hardness grade label to each sea cucumber; And objectively measuring the sample subjected to sensory grading by adopting a texture analyzer, and recording the objective hardness value of the sea cucumber, wherein the data are used for verifying and calibrating the sensory grading result, so that the accuracy of the label is ensured.
6. The method according to claim 1, wherein the determination of the moisture content of the sea cucumber is carried out with reference to a direct drying method in national standard for food safety, determination of moisture in national standard food, GB 5009.3-2016.
7. A method for classifying the hardness of sea cucumber raw materials, which is characterized by comprising the following steps: The sea cucumber raw material hardness classification model constructed by the method according to any one of claims 1-6 is used for determining one or more data of the moisture content, the protein content, the average reflectance value at 960 nm characteristic wavelengths and the glycosaminoglycan leaching amount of a sea cucumber raw material sample to be classified, and inputting the data into the constructed sea cucumber raw material hardness classification model, so that a sea cucumber raw material hardness grade label can be output.
8. A holothurian material hardness grading system, characterized by comprising: The data interface module is used for acquiring the hardness evaluation index data of the sea cucumber raw materials; the label management module is used for supporting hardness evaluation index data of the sea cucumber raw materials and hardness grade label marking to form a training set; the characteristic extraction module is used for extracting one or more characteristic data of the moisture content, the protein content, the average reflectivity value at 960 nm characteristic wavelength and the glycosaminoglycan leaching amount of the sea cucumber raw material sample; The model construction module is used for training a machine learning algorithm according to the corresponding hardness grade label based on the sea cucumber characteristic data to obtain a sea cucumber raw material hardness grading model; the hardness grading module is used for loading the sea cucumber raw material hardness grading model and grading the hardness of the sea cucumber raw material sample to be tested.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, when executing the computer program, being configured to implement the steps of: Acquiring evaluation index data of the hardness of the sea cucumber raw material, wherein the evaluation index data comprise the water content of the sea cucumber, a hardness grade label, an average reflectance value of the sea cucumber at 960 nm characteristic wavelength, the protein content of the sea cucumber and the dissolution amount of the sea cucumber glycosaminoglycan; taking the sea cucumber raw material hardness evaluation index data as a training data set; taking one or more evaluation index data of sea cucumber moisture content, sea cucumber protein content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and sea cucumber glycosaminoglycan leaching amount as training model input data, taking a corresponding hardness grade label as a training model training target, and training a machine learning algorithm to obtain a sea cucumber raw material hardness grading model; and loading the sea cucumber raw material hardness grading model, and grading the hardness of the sea cucumber raw material sample to be tested.
10. A computer readable storage medium, having stored thereon a computer program which, when executed by a processor, is configured to implement the steps of: Acquiring evaluation index data of the hardness of the sea cucumber raw material, wherein the evaluation index data comprise the water content of the sea cucumber, a hardness grade label, an average reflectance value of the sea cucumber at 960 nm characteristic wavelength, the protein content of the sea cucumber and the dissolution amount of the sea cucumber glycosaminoglycan; taking the sea cucumber raw material hardness evaluation index data as a training data set; taking one or more evaluation index data of sea cucumber moisture content, sea cucumber protein content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and sea cucumber glycosaminoglycan leaching amount as training model input data, taking a corresponding hardness grade label as a training model training target, and training a machine learning algorithm to obtain a sea cucumber raw material hardness grading model; and loading the sea cucumber raw material hardness grading model, and grading the hardness of the sea cucumber raw material sample to be tested.

Description

Construction method and application of sea cucumber raw material hardness grading model based on machine learning Technical Field The invention relates to a construction method and application of a sea cucumber raw material hardness grading model based on machine learning, and belongs to the technical field of food processing. Background Sea cucumber is used as a high-value aquatic product, and the raw materials of the sea cucumber need to be strictly classified before processing so as to ensure the uniformity of product quality and specification. However, the quality of sea cucumber raw materials is comprehensively influenced by various complex factors such as growth environment, growth period, salting process and the like, so that the raw materials in different batches and even the same batch have obvious differences in individual size, shape, color and the like. The huge individual variability makes it difficult to process by adopting unified processing procedures, and severely restricts the process of automatic and standardized processing of sea cucumbers. At present, sea cucumber processing enterprises generally adopt a sensory detection method relying on manual experience to classify raw materials, the method mainly relies on naked eye observation and hand feeling touch of operators to judge the grades of sea cucumbers, and has obvious defects that subjective difference is large, judgment standards of different operators are different, judgment of the same operator in different states can possibly fluctuate, classification results are poor in consistency and low in accuracy, and secondly, the standards are difficult to formulate and quantify, sensory experience is difficult to convert into accurate and transportable objective indexes, and stable and unified classification standards cannot be formed. The traditional mode is low in efficiency and high in labor cost, and becomes a key bottleneck for restricting the large-scale and intelligent development of industry. Therefore, aiming at sea cucumber raw materials, an efficient, lossless, objective and accurate grading method is developed to replace the traditional manual sensory grading, so that the intellectualization, automation and standardization of sea cucumber grading are realized, and the method has great significance for improving the whole processing technology level of the sea cucumber industry, guaranteeing the product quality and reducing the production cost, and is also a core target to be realized in the current sea cucumber industry. Disclosure of Invention [ Technical problem ] The existing sea cucumber processing depends on artificial sensory grading, special culture is needed for workers, and the method is long in time consumption and low in efficiency; the invention aims to provide a construction method and application of a sea cucumber raw material hardness grading model based on a machine learning period, which can realize the intellectualization, automation and standardization of sea cucumber grading, and has the advantages of high efficiency, no damage, objectivity and accuracy. Technical scheme In order to achieve the above object, the technical scheme provided is as follows: The first object of the invention is to provide a method for constructing a sea cucumber raw material hardness grading model, which comprises the following steps: acquiring evaluation index data of the hardness of the sea cucumber raw material, wherein the evaluation index data comprise the water content of the sea cucumber, a hardness grade label, an average reflectance value of the sea cucumber at 960 nm characteristic wavelength, the protein content of the sea cucumber and the dissolution amount of the sea cucumber glycosaminoglycan; taking the sea cucumber raw material hardness evaluation index data as a training data set; One or more evaluation index data of sea cucumber moisture content, sea cucumber protein content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and sea cucumber glycosaminoglycan leaching amount are used as training model input data, corresponding hardness grade labels are used as training model training targets, and a machine learning algorithm is trained to obtain a sea cucumber raw material hardness grading model. In one embodiment, the machine learning algorithm is a random forest classification method. In one embodiment, the random forest classification method uses the Scikit-learn library of Python to construct a random forest classifier, sets the number of decision trees to be 100, and defaults the rest parameters. In one embodiment, the training model input data is sea cucumber moisture content, average reflectance value of sea cucumber at 960 nm characteristic wavelength and/or sea cucumber glycosaminoglycan elution. In one embodiment, the hardness grade label is constructed by inviting experienced sea cucumber processing workers to carry out sensory evaluation on the hardness of the sea cucumbe