Wine-making microorganism single colony identification method, system and equipment and storage medium

Abstract

The invention discloses a wine brewing microorganism single colony identification method, system and equipment and a storage medium, and relates to the technical field of microorganism identification. The method comprises the following steps: acquiring and shooting a yeast strain to obtain a bacterial colony image; preprocessing the bacterial colony image to obtain a preprocessed image; inputting the preprocessed image into an improved YOLOv5 neural network for identification and classification to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is realized by using a feature pyramid network and a path aggregation network; and counting various types of microbial colonies to obtain the bacterial colony types, the single colony number and the distribution condition, and generating a report. Various microbial single colonies in the wine brewing process can be accurately identified, classified and counted, so that the efficiency and accuracy of wine brewing quality control are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of microbial identification, and in particular to a method, system, equipment and storage medium for identifying a single colony of brewing microorganisms. Background Art

[0002] In the brewing industry, accurate identification, classification and counting of microorganisms are crucial, especially when determining and controlling the composition of microbial communities during the brewing process. In traditional methods, microscopic observation and culture-based identification are usually used. Existing yeast classification and identification methods mainly include morphological characteristics, physiological and biochemical characteristics, chemical classification methods (coenzyme Q analysis, cell wall component analysis and isozyme electrophoresis analysis) and molecular biological identification methods. These methods have their own advantages and disadvantages:

[0003] The method of morphological characteristics and physiological and biochemical characteristics relies on microscopic observation to identify the morphological and physiological and biochemical characteristics of yeast. Although this method is intuitive and easy to operate, it is extremely time-consuming and requires experienced technicians to operate. In addition, the identification results depend on the operator's experience and judgment, are highly subjective, and cannot accurately distinguish between morphologically similar strains; chemical classification methods identify yeast by analyzing coenzyme Q, cell wall components, and isozyme electrophoresis patterns. These methods can provide more reliable results, but their equipment and reagent costs are high, the operation process is complicated, and they are not suitable for large-scale screening. It requires complex experimental steps, which increases the detection cost and operation difficulty; molecular biological identification methods identify strains through DNA sequence analysis (such as ITS region sequence analysis, PCR-RFLP, etc.). Although this type of method has the advantages of high accuracy and high resolution, it is expensive, and the operation process is complicated and time-consuming. DNA sequencing and analysis require expensive equipment and reagents, and also require operators to have a high level of professional skills. Therefore, there is currently no method that can accurately identify single microbial colonies. Summary of the invention

[0004] The purpose of the present invention is to provide a method, system, equipment and storage medium for identifying single colonies of brewing microorganisms, which can accurately identify, classify and count various single colonies of microorganisms in the brewing process, thereby improving the efficiency and accuracy of brewing quality control.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for identifying a single colony of brewing microorganisms, comprising:

[0007] Collect and photograph yeast strains to obtain colony images;

[0008] Preprocessing the bacterial colony image to obtain a preprocessed image;

[0009] The preprocessed image is input into an improved YOLOv5 neural network for identification and classification to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is achieved by using a feature pyramid network and a path aggregation network;

[0010] Count various types of microbial colonies, obtain the colony types, number of single colonies and distribution, and generate a report.

[0011] Optionally, the process of collecting and photographing the yeast strain is:

[0012] Yeast strains were obtained from the Microbial Resources and Breeding Laboratory of the College of Enology, Northwest Agriculture and Forestry University, and isolated in the laboratory. After being activated with YEPD liquid medium for 12 hours, they were streaked on WLN medium and inverted for culture at 28°C for 5 days. The colony morphology on WLN was observed and pictures were collected. The purified single colony on WLN was picked and activated in YEPD liquid medium for 12 hours. The activated bacterial solution was preserved in a 1.5 mL centrifuge tube with 40% sterile glycerol in a 1:1 ratio. The obtained yeast was subjected to molecular identification of the 26S rDNA D1 / D2 region sequence. At the same time, the preserved yeast was diluted and plated, and images of single colonies of brewing microorganisms were collected. High-resolution camera equipment was used to photograph the single colonies of brewing microorganisms in the culture dish.

[0013] Optionally, the preprocessing includes denoising, contrast enhancement and manual labeling.

[0014] Optionally, the training method of the improved YOLOv5 neural network is:

[0015] Acquire training data; the training data is data and corresponding labels obtained by multiple data enhancement techniques; the multiple data enhancement techniques include rotation, scaling, translation and color transformation of data;

[0016] Build a YOLOv5 pre-trained network;

[0017] The training data is input into the YOLOv5 pre-trained network, and network parameter optimization training is performed with the goal of minimizing the loss between the network output and the data label, and the trained YOLOv5 pre-trained network is determined as the improved YOLOv5 neural network.

[0018] Optionally, it also includes: building a web application; the web application supports image uploading, real-time detection and result display.

[0019] The present invention also provides a brewing microorganism single colony identification system, comprising:

[0020] An image acquisition unit, used to collect and photograph yeast strains to obtain colony images;

[0021] An image processing unit, used for preprocessing the bacterial colony image to obtain a preprocessed image;

[0022] A recognition and classification unit, used for inputting the preprocessed image into an improved YOLOv5 neural network for recognition and classification, so as to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is realized by using a feature pyramid network and a path aggregation network;

[0023] The counting unit is used to count various types of microbial colonies, obtain the colony types, the number of single colonies and their distribution, and generate reports.

[0024] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned brewing microorganism single colony identification method.

[0025] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for identifying a single colony of brewing microorganisms.

[0026] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0027] The present invention discloses a method, system, device and storage medium for identifying single colonies of brewing microorganisms. The method comprises collecting and photographing yeast strains to obtain colony images; preprocessing the colony images to obtain preprocessed images; inputting the preprocessed images into an improved YOLOv5 neural network for identification and classification to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is realized by using a feature pyramid network and a path aggregation network; counting various types of microbial colonies to obtain colony types, single colony numbers and distribution conditions, and generating a report. The present invention can accurately identify, classify and count various microbial single colonies in the brewing process, thereby improving the efficiency and accuracy of brewing quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0029] Figure 1 It is a schematic diagram of the process of the single colony identification method of brewing microorganisms of the present invention;

[0030] Figure 2 Schematic diagram of the data enhancement method structure in the model of this embodiment;

[0031] Figure 3 Schematic diagram of the improved YOLO network model structure in this embodiment;

[0032] Figure 4 Schematic diagram of the structure of the convolutional block attention module of the improved YOLO network model in this embodiment. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] The purpose of the present invention is to provide a method, system, equipment and storage medium for identifying single colonies of brewing microorganisms, which can accurately identify, classify and count various single colonies of microorganisms in the brewing process, thereby improving the efficiency and accuracy of brewing quality control.

[0035] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] like Figure 1-Figure 4 As shown, the present invention provides a method for identifying a single colony of brewing microorganisms, comprising:

[0037] Step 100: Collect and photograph the yeast strain to obtain a colony image.

[0038] Step 200: preprocessing the colony image to obtain a preprocessed image.

[0039] Step 300: Input the preprocessed image into an improved YOLOv5 neural network for identification and classification to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is achieved by using a feature pyramid network and a path aggregation network.

[0040] Step 400: Count various types of microbial colonies to obtain the types of colonies, the number of single colonies and their distribution, and generate a report.

[0041] Based on the above steps, the following specific implementation is provided.

[0042] For the strain identification and image acquisition in the above steps: the yeast strain was obtained from the Microbial Resources and Breeding Laboratory of the College of Wine, Northwest Agriculture and Forestry University. The strain was isolated from the above laboratory, activated with YEPD liquid medium for 12 hours, and then inoculated on WLN medium and inverted at 28°C for 5 days to observe the colony morphology on WLN and collect pictures, and perform preliminary cluster analysis. Pick the purified single colony on WLN and activate it in YEPD liquid medium for 12 hours, and preserve the activated bacterial solution and 40% sterile glycerol in a 1:1 ratio in a 1.5mL centrifuge tube. The obtained yeast was molecularly identified by the 26S rDNA D1 / D2 region sequence, and the preserved yeast was diluted and plated to collect images of single colonies of brewing microorganisms, and high-resolution camera equipment was used to photograph the single colonies of brewing microorganisms in the culture dish.

[0043] For the image preprocessing in the above steps: denoising, contrast enhancement and manual labeling of the collected images. De-noising is used to reduce the noise in the image, contrast enhancement is used to improve the clarity of the image, and manual labeling is performed to ensure the accuracy of the training data. The preprocessed images will be used in the subsequent image segmentation and feature extraction steps.

[0044] For the image segmentation and feature extraction in the above steps: the preprocessed image is processed using the improved YOLOv5 neural network. YOLOv5 is the current advanced target detection model with the characteristics of high efficiency, speed and accuracy.

[0045] This embodiment improves its performance in microbial single colony identification through the following improvements:

[0046] Lightweight network design: Aiming at the characteristics of single colonies of brewing microorganisms, the YOLOv5 model has been lightweighted. The depthwise separable convolutions and pruning technology are used to reduce the number of model parameters and calculations, making it suitable for embedded devices and real-time detection needs.

[0047] Multi-scale feature fusion: The feature pyramid network (FPN) and path aggregation network (PANet) are used to perform multi-scale feature fusion to improve the model's ability to detect colonies of different sizes. By combining low-level and high-level features, the detection accuracy of the model is improved.

[0048] Data enhancement technology: During the training process, a variety of data enhancement techniques are used, such as rotation, scaling, translation, and color transformation, to increase the diversity of training data and improve the generalization ability of the model.

[0049] Colony classification: Use the improved YOLOv5 model to classify single colonies. The classification process uses a pre-trained convolutional neural network (CNN) to extract and classify colonies, which can accurately distinguish different types of microbial colonies.

[0050] Colony counting: Count different types of colonies based on the classification results. Although YOLOv5 does not have a counting function, it is implemented through custom code. Generates statistical data on the number of each colony. These data can be used for further microbial analysis and quality control. The counting function can not only obtain the content of a certain yeast in the wine fermentation process, but also analyze the content of yeast in the grape stock solution, providing a scientific basis for the winemaking process.

[0051] Result output: The classification and counting results are collated and a report is generated. The report includes the type, number and distribution of the colonies in the culture dish. The report can be exported as a PDF or Excel file for further analysis and archiving.

[0052] System deployment: Deploy the trained YOLOv5 network model on the client and make it into a web application (WebApp). Users can upload microbial images through the web application, and the system will classify and count yeast in real time. The web application provides a friendly user interface and supports functions such as image uploading, real-time detection and result display. The system can run on embedded devices or servers to achieve efficient and real-time microbial detection.

[0053] Based on the above-mentioned embodiments, the following examples are provided.

[0054] S1: When performing yeast colony PCR and sequencing, first use a 1.5mL microcentrifuge tube, add 50μL of sterile distilled water (ddH2O), pick a single yeast colony into it and mix it thoroughly. In a 15μL PCR reaction system, add 1μL of template DNA, 1μL of forward primer NL1 (10μM), 1μL of reverse primer NL4 (10μM), 4.5μL of ddH2O and 7.5μL of 2×Rapid Taq MasterMix enzyme. Place the PCR tube in the PCR instrument and set the following reaction program: initial denaturation at 95℃ for 3 minutes; 30-35 cycles of denaturation at 95℃ for 15 seconds, annealing at 50℃ for 15 seconds and extension at 72℃ for 15 seconds; finally, extend at 72℃ for 5 minutes, and store at 4℃ after the reaction. Analyze the PCR product using agarose gel electrophoresis. After the specific fragment is detected, the PCR product is sent to Qingke Bio for sequencing analysis.

[0055] Among them, WLN (Wallerstein laboratory nutrient agar) medium is a selective medium widely used for yeast and bacterial culture, especially in the process of winemaking and industrial fermentation. Different yeasts have different absorption abilities for bromocresol green in WLN medium, so that the colonies show different colors, so that different species of yeast can be distinguished. The shape, size and color of the yeast are the main basis for judging the species of yeast colonies.

[0056] The preserved yeast cultures of different species were diluted so that at least 10 yeast colonies could grow on the diluted WLN medium. After the dilution multiple was determined, the culture plates were plated and then placed upside down in a 28°C constant temperature incubator for 5 days.

[0057] S2: Collect photos of yeast colonies that have grown on WLN medium for 5 days, at least 100 photos of each yeast, perform data enhancement on the photos, manually annotate the yeast colonies in the enhanced photos, and generate a data set. For all the photos in the data set, this embodiment divides them into training set, validation set and test set in a ratio of 8:1:1. The training set is used to input into the neural network model for model training to ensure that the model can learn as many features and patterns as possible; the validation set is used to verify the output results of the training, evaluate the model performance and prevent overfitting; finally, the test set is used to evaluate the performance of the model to ensure the reliability and representativeness of the model evaluation results. The yeast database includes different species of yeast of different genera.

[0058] S3: construct a yeast classification neural network model, use the training set to train the neural network model, and the neural network model adopts an improved YOLO network model.

[0059] S4: Input the image-enhanced implementation photo into the trained neural network model to obtain the categories and quantity of yeast on the culture dish.

[0060] In step S2, the image enhancement technology is mainly Mosaic Augmentation, which generates a complex background image containing multiple image information by randomly scaling and splicing four images together. Mosaic enhancement technology effectively enriches the detection data set, adds many small targets, and enhances the robustness of the network model. By processing the data of the fourth picture, the size of each batch can be reduced while maintaining the detection effect, which helps to improve the training speed and accuracy of the model during the training process, and reduces the demand for GPU video memory resources during the detection process. In addition, it also uses data enhancement methods such as random cropping, color jittering, translation, scaling and rotation. These technologies expand the scale of the training set through randomized image changes, reduce the model's dependence on specific features, and improve the model's generalization ability.

[0061] As a system structure for realizing the above technical method, it includes:

[0062] Image acquisition unit: high-resolution camera, light source and its bracket.

[0063] Image processing unit: A computer or embedded device equipped with image processing software and algorithms.

[0064] Model training unit: computing resources used for model training and optimization.

[0065] Identification of taxonomic units and technical units: used to identify and count single colonies.

[0066] User interface unit: Web application (WebApp), used for user operation and result display.

[0067] The functions of the above system are:

[0068] Image acquisition: The image acquisition unit is used to obtain high-quality images of single colonies of brewing microorganisms and adjust the position and angle of the camera and light source to obtain the best shooting effect.

[0069] Image preprocessing: The image processing unit performs denoising, contrast enhancement and manual labeling on the collected images to ensure the accuracy and completeness of the training data.

[0070] Model training and optimization: The model training unit uses lightweight design, multi-scale feature fusion and data enhancement technology to train and optimize the improved YOLOv5 neural network. The model can accurately segment and extract the features of a single colony with low computing resources.

[0071] Real-time detection and counting: Users upload microbial images through the web application, and the system classifies and counts yeasts in real time, and generates charts and reports of classification and counting results for easy analysis and archiving.

[0072] As can be seen from the above, compared with the existing yeast colony identification technology, this embodiment has the following technical effects:

[0073] 1. The first application of the YOLO model for single colony detection of brewing microorganisms: This paper innovatively applies the YOLOv5 model to the detection of single colonies of brewing microorganisms. Through lightweight improvements and multi-scale feature fusion, the detection speed and accuracy of the model are improved. This is the first time that the YOLO model has been used in brewing microorganism detection, filling the technical gap in this field.

[0074] 2. Lightweight improvement: In view of the characteristics of single colonies of brewing microorganisms, the YOLOv5 model has been lightweighted. Depthwise Separable Convolutions and Pruning techniques are used. The Depthwise Separable Convolutions reduces the number of parameters and computational complexity of each layer of convolution operations in the model while maintaining the model's feature extraction capabilities. The Pruning technique optimizes the trained model, removes redundant network connections and parameters, and further reduces the model's complexity and computational requirements, enabling the model to run efficiently on embedded devices. This improvement solves the problem of high computational resource consumption in traditional deep learning models and achieves efficient real-time detection.

[0075] 3. Multi-scale feature fusion: In the YOLOv5 model, the feature pyramid network (FPN) and path aggregation network (PANet) are used for multi-scale feature fusion, which improves the model's ability to detect colonies of different sizes. By combining low-level and high-level features, the detection accuracy of the model is improved. This technological innovation enables the model to accurately identify microbial colonies in complex backgrounds.

[0076] 4. Adopt CBAM module: The convolutional block attention module (CBAM) is integrated into the YOLO model. By modeling the spatial attention and channel attention of the feature map, the model's ability to pay attention to the characteristics of a single microbial colony is enhanced, thereby improving the accuracy and robustness of detection.

[0077] 5. Customized counting function: Although the YOLOv5 model does not have a built-in counting function, the present invention implements the counting function of classified colonies through customized code. According to the type and position of the classified colonies, the proportion of each type of colony in a batch of images is calculated, and combined with the number and growth of each type of colony in the winemaking process, a scientific basis is provided for the microbial control of the winemaking process.

[0078] 6. Data enhancement technology: In order to enhance the generalization ability of the model, a variety of data enhancement technologies were introduced during the training process, including rotation, scaling, translation, and color transformation. These data enhancement technologies increase the diversity of training data and help the model better adapt to microbial single colony images under different shooting conditions.

[0079] 7. System deployment and real-time detection: Deploy the trained YOLOv5 model on embedded devices and develop a web application (WebApp) to enable users to upload images through the web page for real-time detection. The web application provides a friendly user interface and supports functions such as image uploading, real-time detection and result display. The system can run on embedded devices or servers to ensure efficient and real-time microbial detection capabilities. This innovation not only realizes efficient and automated brewing microbial detection, but also provides a friendly user interface for convenient user operation and data management.

[0080] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0081] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for identifying a single colony of brewing microorganisms, characterized in that: include: Collect and photograph yeast strains to obtain colony images; Preprocessing the bacterial colony image to obtain a preprocessed image; The preprocessed image is input into an improved YOLOv5 neural network for identification and classification to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is achieved by using a feature pyramid network and a path aggregation network; Count various types of microbial colonies, obtain the colony types, number of single colonies and distribution, and generate a report.

2. The method for identifying a single colony of brewing microorganisms according to claim 1, characterized in that: The process of collecting and photographing the yeast strain is as follows: Yeast strains were obtained from the Microbial Resources and Breeding Laboratory of the College of Enology, Northwest Agriculture and Forestry University, and isolated in the laboratory. After being activated with YEPD liquid medium for 12 hours, they were streaked on WLN medium and inverted for culture at 28°C for 5 days. The colony morphology on WLN was observed and pictures were collected. The purified single colony on WLN was picked and activated in YEPD liquid medium for 12 hours. The activated bacterial solution was preserved in a 1:1 ratio with 40% sterile glycerol in a 1.5 mL centrifuge tube. The obtained yeast was subjected to molecular identification of the 26SrDNAD1 / D2 region sequence. At the same time, the preserved yeast was diluted and plated, and images of single colonies of brewing microorganisms were collected. High-resolution camera equipment was used to photograph the single colonies of brewing microorganisms in the culture dish.

3. The method for identifying a single colony of brewing microorganisms according to claim 1, characterized in that: The preprocessing includes denoising, contrast enhancement and manual labeling.

4. The method for identifying a single colony of brewing microorganisms according to claim 1, characterized in that: The training method of the improved YOLOv5 neural network is: Acquire training data; the training data is data and corresponding labels obtained by multiple data enhancement techniques; the multiple data enhancement techniques include rotation, scaling, translation and color transformation of data; Build a YOLOv5 pre-trained network; The training data is input into the YOLOv5 pre-trained network, and network parameter optimization training is performed with the goal of minimizing the loss between the network output and the data label, and the trained YOLOv5 pre-trained network is determined as the improved YOLOv5 neural network.

5. The method for identifying a single colony of brewing microorganisms according to claim 1, characterized in that: Also includes: Build web applications; The web application supports image uploading, real-time detection and result display.

6. A single colony identification system for brewing microorganisms, characterized in that: include: An image acquisition unit, used to collect and photograph yeast strains to obtain colony images; An image processing unit, used for preprocessing the bacterial colony image to obtain a preprocessed image; A recognition and classification unit, used for inputting the preprocessed image into an improved YOLOv5 neural network for recognition and classification, so as to obtain different types of microbial colonies; the improved YOLOv5 neural network is constructed based on deep separable convolution, pruning and multi-scale feature fusion; the multi-scale feature fusion is realized by using a feature pyramid network and a path aggregation network; The counting unit is used to count various types of microbial colonies, obtain the colony types, the number of single colonies and their distribution, and generate reports.

7. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the brewing microorganism single colony identification method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements the brewing microorganism single colony identification method as described in any one of claims 1 to 5.

Images