A real-time detection method and system for the waist explosion rate of rice based on image recognition

Abstract

The present invention discloses a real-time detection method and system for the waist explosion rate of rice based on image recognition, which relates to image optimization technology and is used to improve the problem of untimely detection caused by the change of the waist explosion rate due to the degree of association between internal and external factors of rice. It includes obtaining rice image samples, recording the waist explosion areas of rice grains and calculating the morphological change rate, collecting the surface crack density and the change rate of grain volume of rice grains in the rice image samples to evaluate the damage degree of rice grains, generating a marked detection time interval by using the morphological change rate and the damage degree of rice grains, screening the rice image samples according to the detection time interval, counting the proportion of the waist explosion area for waist explosion rating. When the waist explosion rating is low, collecting the ventilation data of the rice storage environment and analyzing the ventilation fluctuation, evaluating the internal stress distribution by combining the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area, and selecting different algorithm models according to the internal stress distribution of rice grains to predict the change trend of the waist explosion rate.

Description

Technical Field

[0001] The present invention relates to the technical field of image optimization. More specifically, the present invention relates to a real-time detection method and system for the waist explosion rate of rice based on image recognition. Background Art

[0002] With the development of agricultural intelligence and the improvement of processing quality requirements, the detection technology of the waist explosion rate of rice has gradually attracted attention. The application of image optimization technology in the detection of the waist explosion rate of rice can give timely warnings, thereby lengthening the response time for dealing with waist explosion.

[0003] The existing technology has the following deficiencies:

[0004] Traditional methods for detecting the waist explosion rate of rice mainly rely on manual visual inspection or simple mechanical screening. Although these methods can provide certain reference data, they have the characteristics of low efficiency, strong subjectivity, and difficulty in realizing continuous detection. In addition, the existing technology usually only evaluates the overall appearance of rice and fails to fully focus on the accurate quantification of the key index of the waist explosion rate, and does not optimize the subtle morphological changes of static rice grains, resulting in low detection efficiency of the waist explosion rate of rice. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the existing technology, the embodiments of the present invention provide a real-time detection method and system for the waist explosion rate of rice based on image recognition. By calculating the morphological change rate of the waist explosion area, generating the optimal detection time interval, evaluating the influence of internal and external factors, and predicting the change trend of the waist explosion rate, different algorithm models are selected to predict the waist explosion rate of rice to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A real-time detection method for the waist explosion rate of rice based on image recognition, comprising the following steps:

[0008] Step S1: Obtain multiple rice image samples, record the waist explosion area of rice grains according to the edge detection algorithm and calculate the morphological change rate, and collect the surface crack density and the grain volume change rate of rice grains in each rice image sample to evaluate the damage degree of rice grains;

[0009] Step S2: Generate a marked detection time interval by using the morphological change rate and the damage degree of rice grains, screen the rice image samples according to the detection time interval, and count the proportion of the waist explosion area for waist explosion rating; Step S3: When the waist explosion rating is low, collect the ventilation data of the rice storage environment and analyze the ventilation fluctuation, and evaluate the waist explosion rate caused by the internal stress distribution in combination with the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area;

[0010] Step S4: Analyze the correlation degree between internal and external factors of rice grains based on the waist explosion rate caused by the internal stress distribution of rice grains, and select different algorithm models to predict the changing trend of the waist explosion rate.

[0011] In a preferred embodiment, in step S1, the steps of calculating the morphological change rate of the waist explosion area include: detecting the area of the waist explosion area at two time points with different time intervals randomly set in each analysis sample, subtracting the area of the waist explosion area detected at the latter time point from the area of the waist explosion area detected at the former time point, dividing the obtained calculation result by the corresponding time interval to obtain the morphological change rate of the waist explosion area, merging the morphological change rates of the waist explosion areas calculated for each analysis sample into a change rate data set, and at the same time merging the time intervals into a detection time data set.

[0012] In a preferred embodiment, in step S1, the surface crack density of rice grains is the number of cracks per unit area, the grain volume change rate is the ratio of the volume of rice grains at the detection time point to the initial volume of rice grains, and the sum of the normalized surface crack density of rice grains and the grain volume change rate is used as the damage degree of rice grains, and the damage degrees of rice grains in each analysis sample are merged into a damage degree data set.

[0013] In a preferred embodiment, in step S2, when generating a label for the detection time interval using the morphological change rate and damage degree of rice grains, the least absolute shrinkage and selection operator is used for processing. The specific steps are as follows:

[0014] Integrate the three input features of each rice image sample into a feature matrix, and the sparsity constraint can be set as the L1 norm penalty term: , and the loss function is: , is the true value of the input feature, is the model output value, n is the number of feature matrices, p is the number of input features in the feature matrix, is the regression coefficient of the model, and λ is the regularization parameter used to control the degree of sparsity.

[0015] In a preferred embodiment, in step S2, after generating a label for the detection time interval, compare the proportion of the waist explosion area in the rice image sample under the label detection time interval with a preset waist explosion proportion threshold to classify the rice image sample as having a high waist explosion rating or a low waist explosion rating.

[0016] In a preferred embodiment, in step S3, the ventilation data of the rice storage environment is the ventilation rate of the rice storage environment. Set multiple equal time periods to count the ventilation rate of the rice storage environment, and use the standard deviation of the ventilation rate in each time period as the ventilation fluctuation of the rice storage environment.

[0017] In a preferred embodiment, in step S3, the ventilation fluctuations of the rice storage environment at different times are selected to calculate the average value as the ventilation fluctuation benchmark of the rice storage environment;

[0018] Mark the rice storage environment collection time when the ventilation fluctuation is 0, record the waist explosion rate of the rice grains at the marked collection time, calculate the external loss waist explosion rate by combining the waist explosion rate of the rice grains under the ventilation fluctuation benchmark, calculate the external loss ratio by integrating the ventilation data at the time of rice image sample collection and the ventilation fluctuation benchmark. Finally, take the product of the external loss waist explosion rate and the external loss ratio as the external loss waist explosion rate of the grains in the rice image sample;

[0019] Subtract the grain damage rate of the rice grains from the external loss waist explosion rate to obtain the waist explosion rate caused by the internal stress distribution of the rice grains.

[0020] In a preferred embodiment, in step S4, when the difference between the external loss waist explosion rate of the rice grains and the waist explosion rate caused by the internal stress distribution is within a preset correlation threshold, it is determined that the internal and external factor correlation relationship of the rice grains is strong; otherwise, it is determined that the internal and external factor correlation relationship of the rice grains is weak.

[0021] In a preferred embodiment, in step S4, when the internal and external factor correlation relationship of the rice grains in the rice image sample is weak, the random forest algorithm in the decision tree model is used to predict the change trend of the waist explosion rate of the rice grains in the rice image sample; otherwise, the convolutional neural network model under the deep learning framework is used for prediction.

[0022] A real-time detection system for the waist explosion rate of rice based on image recognition, which is used to implement the above-mentioned real-time detection method for the waist explosion rate of rice based on image recognition, includes an image acquisition module, a sample analysis module, an environment evaluation module, and a prediction optimization module;

[0023] The image acquisition module is used to acquire rice image samples on the plane of the sample to be measured;

[0024] The sample analysis module is used to extract the contour information of the rice grains and remove noise interference, and calculate the area of the waist explosion area;

[0025] The environment evaluation module is used to collect the ventilation data of the rice storage environment;

[0026] The prediction optimization module is used to select a suitable algorithm model according to the internal and external factor correlation relationship and output the waist explosion rate prediction result.

[0027] The technical effects and advantages of the real-time detection method and system for the waist explosion rate of rice based on image recognition of the present invention:

[0028] The present invention obtains multiple rice image samples, records the waist explosion areas of rice grains and calculates the morphological change rate, collects the surface crack density and the grain volume change rate of rice grains in each rice image sample to evaluate the damage degree of rice grains, generates a marked detection time interval by using the morphological change rate and the damage degree of rice grains, screens the rice image samples corresponding to the detection time interval according to the damage of rice grains, improves the analysis reliability of image data, facilitates subsequent algorithm model prediction and selection processing, screens the rice image samples according to the detection time interval, and counts the proportion of the waist explosion area for waist explosion rating. When the waist explosion rating is low, the ventilation data of the rice storage environment is collected and the ventilation fluctuation is analyzed. The lower the waist explosion rating of the rice is, the more attention needs to be paid to the rice grains in the corresponding image samples. Combining the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area to evaluate the internal stress distribution, and selecting different algorithm models according to the internal stress distribution of rice grains to predict the change trend of the waist explosion rate, thereby improving the accuracy of rice waist explosion identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of a real-time detection method for the rice waist explosion rate based on image recognition according to the present invention.

[0030] Figure 2 It is a flowchart of a real-time detection system for the rice waist explosion rate based on image recognition according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] The present invention obtains multiple rice image samples, records the waist explosion areas of rice grains and calculates the morphological change rate, collects the surface crack density and the grain volume change rate of rice grains in each rice image sample to evaluate the damage degree of rice grains, generates a marked detection time interval by using the morphological change rate and the damage degree of rice grains, screens the rice image samples corresponding to the detection time interval according to the damage of rice grains, improves the analysis reliability of image data, facilitates subsequent algorithm model prediction and selection processing, screens the rice image samples according to the detection time interval, and counts the proportion of the waist explosion area for waist explosion rating. When the waist explosion rating is low, the ventilation data of the rice storage environment is collected and the ventilation fluctuation is analyzed. Combining the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area to evaluate the internal stress distribution, and selecting different algorithm models according to the internal stress distribution of rice grains to predict the change trend of the waist explosion rate, thereby improving the accuracy of rice waist explosion identification.

[0033] Example 1, a real-time detection method for the rice waist explosion rate based on image recognition, as Figure 1As shown in the figure, it includes the following steps:

[0034] Step S1: Obtain multiple rice image samples, record the waist explosion area of rice grains according to the edge detection algorithm and calculate the morphological change rate, and collect the surface crack density and the particle volume change rate of rice grains in each rice image sample to evaluate the damage degree of rice grains;

[0035] Step S2: Generate a marked detection time interval by using the morphological change rate and the damage degree of rice grains, and screen the rice image samples according to the detection time interval, and count the proportion of the waist explosion area for waist explosion rating; Step S3: When the waist explosion rating is low, collect the ventilation data of the rice storage environment and analyze the ventilation fluctuation, and evaluate the waist explosion rate caused by the internal stress distribution in combination with the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area;

[0036] Step S4: Analyze the correlation degree between internal and external factors of rice grains according to the waist explosion rate caused by the internal stress distribution of rice grains and select different algorithm models to predict the change trend of the waist explosion rate.

[0037] The specific implementation is as follows:

[0038] In step S1, when obtaining rice image samples, select equal amounts of rice from the same processing batch as analysis samples, randomly set two time points at different time intervals to detect the morphological changes of the waist explosion area, obtain the rice grain images on multiple sample planes to be measured through an industrial camera, process the collected images using the edge detection algorithm and extract the contour information of the rice grains, perform denoising processing on the contour information of the rice grains, calculate the area of the waist explosion area of the rice grain samples according to the denoised contour information, and record the waist explosion area values at the two time points. By subtracting the waist explosion area detected at the latter time point from the area at one time point and then dividing by the corresponding time interval, the morphological change rate of the waist explosion area is obtained.

[0039] Merge the morphological change rates of the waist explosion areas calculated for each analysis sample into a change rate data set, and at the same time merge the time intervals into a detection time data set. Use the midpoint of the time interval of each rice image sample as the detection time point for the surface crack density and the particle volume change rate of the rice grains and collect them.

[0040] Furthermore, for the image sequences of each rice grain image sample at different detection time points, extract the waist explosion area through the image segmentation algorithm. At the same time, use the midpoint of the time interval between every two adjacent image time points as the detection time point to construct a detection time data set.

[0041] Among them, the image segmentation algorithm can include but is not limited to methods based on threshold segmentation, methods based on edge detection, methods based on region growing, and deep learning-based image semantic segmentation algorithms; the above algorithms can be flexibly selected or combined according to the actual image quality and particle edge characteristics to ensure the accuracy and stability of rice contour extraction.

[0042] The surface crack density of rice grains is obtained by counting the number of cracks within a unit area, and the particle volume change rate is the ratio of the volume of rice grains at the detection time point to the initial volume of rice grains. The sum of the normalized surface crack density of rice grains and the particle volume change rate is used as the damage degree of rice grains, and the damage degrees of rice grains in each analysis sample are combined into a damage degree dataset.

[0043] Among them, the surface crack density of rice grains refers to the number of cracks on the surface of rice grains per unit area, specifically used to quantitatively characterize the distribution of surface cracks of rice grains. The larger the value, the more structural damage exists on the particle surface, which may affect its integrity during processing, transportation, or cooking.

[0044] It should be noted that the normalization methods for the surface crack density of rice grains and the particle volume change rate are not unique. For example, using the Min - Max normalization method to process the surface crack density of rice grains and the particle volume change rate, the larger the surface crack density or volume change rate of rice grains, the higher the damage degree of rice grains.

[0045] The particle volume change rate refers to the change ratio of the volume of rice grains at the detection time point relative to the initial volume, which is calculated by taking the ratio of the estimated volume of the particle at the current moment to the volume of the particle at the initial moment to obtain the particle volume change rate.

[0046] Furthermore, the particle volume change rate is obtained by using a high - resolution image acquisition system or three - dimensional reconstruction technology to acquire the morphological data of rice grains at different time points, and using an image segmentation algorithm to extract the contour information of the particles; then, according to the particle contour, the three - dimensional volume of the particle is calculated by volume estimation methods, such as based on the rotational projection method or multi - view stereo reconstruction technology, to obtain the particle volume at different time points; then, taking the particle volume at the initial moment as a reference, the particle volume change rate is calculated.

[0047] Specifically, for the setting of time points, the experimenters can comprehensively determine it according to actual factors such as the number of samples, sampling frequency, equipment response speed, and analysis requirements, which will not be elaborated here.

[0048] In step S2, the least absolute shrinkage and selection operator is used to screen each rice image sample and generate a marked detection time interval, and the proportion of the waist explosion area in the rice image sample under the marked detection time interval is statistically calculated. The specific steps are as follows:

[0049] Taking the waist explosion area morphological change rate dataset, the detection time dataset, and the damage degree dataset as input features, adding a sparsity constraint through the loss function to set the influence weight coefficient, using the coordinate descent method to solve the weight coefficient, and taking the output detection time interval as the marked time interval.

[0050] In the least absolute shrinkage and selection operator, the three input features of each rice image sample are integrated into a feature matrix, and the sparsity constraint can be set as the L1 norm penalty term: , and the loss function is: , is the true value of the input feature, is the model output value, n is the number of feature matrices, p is the number of input features in the feature matrix, is the regression coefficient of the model, and λ is the regularization parameter used to control the degree of sparsity.

[0051] When rating the waist explosion of rice grains, compare the proportion of the waist explosion area in the rice image sample under the marked detection time interval with the preset waist explosion proportion threshold. When the proportion of the waist explosion area in the rice image sample under the detection time interval is lower than the waist explosion proportion threshold, it is judged that the waist explosion rating of the rice image sample is high; when the proportion of the waist explosion area in the rice image sample under the detection time interval exceeds the waist explosion proportion threshold, it is judged that the waist explosion rating of the rice image sample is low.

[0052] It should be noted that solving the weight coefficient is used to integrate the three input features in this example into a feature matrix. The coordinate descent method is an optimization algorithm that finds the minimum or maximum value of a function by gradually optimizing each coordinate. In this example, it is used to calculate the optimal weight coefficient of the three input features; the above-mentioned screening of the marked detection time interval requires the central processing unit to complete data processing and model training, and the prediction results are used to guide subsequent detection operations.

[0053] In step S3, when the waist explosion rating of the rice image sample is low, collect the ventilation data of the rice storage environment. The ventilation data of the rice storage environment is the ventilation rate of the rice storage environment. Set multiple equal time periods to statistically calculate the ventilation rate of the rice storage environment and calculate the standard deviation of the ventilation rate as the ventilation fluctuation of the rice storage environment;

[0054] Evaluate the influence of the external storage environment of rice on the waist explosion rate of rice according to the ventilation fluctuation of the rice storage environment. Select different periods to calculate the average value of the ventilation fluctuation of the rice storage environment as the ventilation fluctuation benchmark of the rice storage environment, and detect the waist explosion rate of rice under the ventilation fluctuation benchmark;

[0055] Among them, the detection method of the waist explosion rate of rice under the ventilation fluctuation benchmark includes links such as storage environment parameter collection, sample image processing, and waist explosion recognition and statistics, which can effectively quantify the influence of the external ventilation environment on the structural stability of rice, and is applicable to application scenarios such as rice storage stability evaluation, warehousing optimization control, and quality prediction;

[0056] Furthermore, within the time period corresponding to the ventilation fluctuation benchmark, randomly select representative rice samples from the rice storage batch, use a high-resolution image acquisition system to obtain sample images, and adopt image preprocessing and image segmentation algorithms to identify the waist explosion characteristics in the rice grains. By statistically calculating the ratio between the number of grains with waist explosion and the total number of grains, calculate the waist explosion rate of rice under the current ventilation fluctuation benchmark, and then use it to evaluate the influence of the external ventilation environment on the structural stability of rice;

[0057] Specifically, the selection of different periods can be based on the natural stage division of the storage cycle, seasonal changes, the operating state of ventilation equipment, the time points of environmental parameter mutations, etc., which will not be elaborated here;

[0058] Mark the acquisition time of the rice storage environment when the ventilation fluctuation is 0, record the waist explosion rate of the rice grains at the acquisition time, subtract the waist explosion rate of the rice grains at the acquisition time from the waist explosion rate of the rice grains under the ventilation fluctuation benchmark to obtain the external damage waist explosion rate, call the ventilation data of the rice storage environment during the acquisition of the rice image sample and calculate the ventilation fluctuation, and use the ratio of the ventilation fluctuation to the ventilation fluctuation benchmark as the external damage ratio. Multiply the external damage waist explosion rate by the external damage ratio to obtain the external damage waist explosion rate of the rice grains in the rice image sample;

[0059] During the acquisition time, count the number of rice grains with a volume lower than the preset volume threshold, and use the ratio of it to the total number of grains as the grain damage rate of the rice grains;

[0060] Subtract the grain damage rate of the rice grains from the external damage waist explosion rate to obtain the waist explosion rate caused by the internal stress distribution of the rice grains, and record it.

[0061] It should be noted that the waist explosion rate of rice grains is composed of the external damage waist explosion rate of rice grains and the waist explosion rate caused by the internal stress distribution. The correlation between the two can also be quantified by other correlation analysis methods, providing a basis for the selection of subsequent prediction models.

[0062] In step S4, different algorithm models are selected according to the internal stress distribution of rice grains to predict the change trend of the waist explosion rate. The external damage waist explosion rate of rice grains is compared with the waist explosion rate caused by the internal stress distribution. The closer the external damage waist explosion rate of rice grains is to the waist explosion rate caused by the internal stress distribution, the stronger the correlation between the internal and external factors of rice grains. Because when the external damage waist explosion rate of rice grains is close to the waist explosion rate caused by the internal stress distribution, it is easier to produce damaging interactions.

[0063] When the difference between the external damage waist explosion rate of rice grains and the waist explosion rate caused by the internal stress distribution is within the preset correlation threshold, it is determined that the correlation between the internal and external factors of rice grains is strong; otherwise, it is determined that the correlation between the internal and external factors of rice grains is weak.

[0064] When the correlation between the internal and external factors of rice grains in the rice image sample is weak, the random forest algorithm in the decision tree model is used to predict the change trend of the waist explosion rate of rice grains in the rice image sample; when the correlation between the internal and external factors of rice grains in the rice image sample is strong, the convolutional neural network model under the deep learning framework is used for prediction.

[0065] Among them, the random forest algorithm in the decision tree model is an ensemble learning method, belonging to the ensemble classification and regression algorithm based on decision trees. This algorithm improves the overall prediction performance and robustness by constructing multiple decision tree models and integrating the prediction results of each tree through voting (classification task) or averaging (regression task).

[0066] The prediction result is output to the user interface for the operator's reference. The random forest algorithm obtains the final prediction value by constructing multiple decision trees and voting on the prediction results of each tree, while the convolutional neural network model extracts image features through multiple convolutional and pooling operations and completes the prediction task.

[0067] It should be noted that the selection basis of the two algorithm models is the strength of the correlation between internal and external factors to ensure the accuracy and reliability of the prediction results.

[0068] Embodiment 2, a real-time detection method for the waist explosion rate of rice based on image recognition, as Figure 2 shown, includes the following modules:

[0069] Image acquisition module: used to acquire the rice image sample on the plane of the sample to be measured.

[0070] Sample analysis module: used to extract the contour information of rice grains, remove noise interference, and calculate the area of the waist explosion area.

[0071] Environment assessment module: used to acquire the ventilation data of the rice storage environment.

[0072] Prediction Optimization Module: It is used to select an appropriate algorithm model according to the correlation between internal and external factors and output the prediction result of the waist explosion rate.

[0073] In practical applications, this example gives a usage method, and the specific implementation is as follows:

[0074] First, evenly spread the rice samples to be measured on the plane of the samples to be measured, ensuring that the surface of the samples is flat and there is no obvious accumulation;

[0075] The industrial camera is installed on a fixed bracket and set at a 45-degree angle to the plane of the samples to be measured to avoid the influence of specular reflection on the image acquisition quality;

[0076] The uniform light source is located directly above the plane of the samples to be measured. Using an LED lamp bead array design, it can evenly cover the surface of the samples, thereby reducing the image noise interference caused by uneven illumination;

[0077] After starting the system, the industrial camera acquires high-definition images of the rice samples and transmits the image data to the sample analysis module for processing;

[0078] In the sample analysis module, the Canny algorithm is specifically used in the edge detection algorithm to extract the contour information of the rice grains; since the image acquisition process may be interfered by ambient light or tiny particles on the sample surface, noise points are removed through opening and closing operations, thereby improving the accuracy of contour extraction;

[0079] Subsequently, based on the denoised contour information, the waist explosion area calculation module calculates the area of the waist explosion area using the pixel counting method and records the area values of the waist explosion area at two time points; by subtracting the area of the waist explosion area at the latter time point from the area at the previous time point and dividing by the corresponding time interval, the morphological change rate of the waist explosion area is obtained;

[0080] The key to this step is to ensure that the calculation accuracy of the waist explosion area reaches the single-pixel level through a high-resolution industrial camera and precise algorithm design.

[0081] The central processing unit receives the dataset of the morphological change rate of the waist explosion area, the dataset of the detection time, and the dataset of the damage degree, and selects the best detection time interval through the least absolute shrinkage and selection operator;

[0082] Specifically, a sparsity constraint is added to the loss function, the influence weight coefficient is set to eliminate redundant features, and the coordinate descent method is used to solve the optimal weight coefficient; the core principle of this process is to reduce the model complexity through the sparsity constraint, thereby improving the prediction efficiency and accuracy; finally, the system determines the best detection time interval according to the screening results and applies it to subsequent sample analysis.

[0083] The system compares the actual proportion of the waist explosion area with a preset threshold to obtain the waist explosion rating of the rice image sample; if the proportion of the waist explosion area is low, it collects the ventilation data of the rice storage environment, and the ventilation data of the rice storage environment is the ventilation rate of the rice storage environment, and evaluates the impact of the external environment on particle damage, that is, the external damage waist explosion rate;

[0084] Finally, analyze the internal and external factor correlation relationship based on the external damage waist explosion rate and the waist explosion rate caused by the internal stress of the rice grains, and then use the internal and external factor correlation relationship of the rice grains to select different algorithm models to predict the change trend of the waist explosion rate.

[0085] Through the above module division of labor and cooperation, a technical solution combining image recognition and multi-factor analysis is realized, which not only meets the requirements of modern agricultural production for intelligent quality inspection, but also provides a new technical path for rice processing quality control.

[0086] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0087] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and invention constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0088] In addition, in each embodiment of the present application, the functional modules can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0089] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0090] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A real-time detection method for the waist explosion rate of rice based on image recognition, characterized in that, Including the following steps, Step S1: Obtain multiple rice image samples, record the waist explosion areas of rice grains according to the edge detection algorithm and calculate the morphological change rate, and collect the surface crack density and grain volume change rate of rice grains in each rice image sample to evaluate the damage degree of rice grains; Step S2: Generate a marked detection time interval by using the morphological change rate and damage degree of rice grains, screen the rice image samples according to the detection time interval, and count the proportion of the waist explosion area for waist explosion rating; Step S3: When the waist explosion rating is low, collect the ventilation data of the rice storage environment and analyze the ventilation fluctuation, and evaluate the waist explosion rate caused by the internal stress distribution by combining the ventilation fluctuation of the rice storage environment and the morphological change rate of the waist explosion area; Step S4: Analyze the correlation degree between internal and external factors of rice grains according to the waist explosion rate caused by the internal stress distribution of rice grains and select different algorithm models to predict the change trend of the waist explosion rate; In step S2, set a marked detection time interval before the waist explosion rating. Compare the proportion of the waist explosion area in the rice image sample under the marked detection time interval with the preset waist explosion proportion threshold. When the proportion of the waist explosion area in the rice image sample under the detection time interval is lower than the waist explosion proportion threshold, it is determined that the waist explosion rating of the rice image sample is high; When the proportion of the waist explosion area in the rice image sample under the detection time interval exceeds the waist explosion proportion threshold, it is determined that the waist explosion rating of the rice image sample is low; In step S3, the ventilation data of the rice storage environment is the ventilation rate of the rice storage environment. Set multiple equal time periods to count the ventilation rate of the rice storage environment, and use the standard deviation of the ventilation rate in each time period as the ventilation fluctuation of the rice storage environment; In step S3, select different periods to calculate the average value of the ventilation fluctuation of the rice storage environment as the ventilation fluctuation reference of the rice storage environment; Mark the collection time of the rice storage environment when the ventilation fluctuation is 0, record the waist explosion rate of the rice grains at the marked collection time, calculate the external damage waist explosion rate by combining the waist explosion rate of the rice grains under the ventilation fluctuation reference, calculate the external damage ratio by integrating the ventilation data at the time of collecting the rice image sample and the ventilation fluctuation reference. Finally, use the product of the waist explosion rate of the rice grains under the ventilation fluctuation reference and the external damage ratio as the external damage waist explosion rate of the grains in the rice image sample; Subtract the grain damage rate of the rice grains from the external damage waist explosion rate to obtain the waist explosion rate caused by the internal stress distribution of the rice grains; The waist explosion rate of rice grains is the ratio of the number of grains with waist explosion to the total number of grains during the collection time, When calculating the grain damage rate of rice grains, count the number of rice grains with a volume lower than the preset volume threshold during the collection time, and use the ratio of it to the total number of grains as the grain damage rate of rice grains.

2. A real-time detection method for the waist explosion rate of rice based on image recognition according to claim 1, characterized in that: In step S1, the steps for calculating the morphological change rate of the waist explosion area include: randomly setting two time points with different time intervals in each analysis sample to detect the area of the waist explosion area, subtracting the area of the waist explosion area detected at the latter time point from the area of the waist explosion area detected at the former time point, dividing the obtained calculation result by the corresponding time interval to obtain the morphological change rate of the waist explosion area, combining the morphological change rates of the waist explosion area calculated for each analysis sample into a change rate data set, and at the same time combining the time intervals into a detection time data set.

3. A real-time detection method for the waist explosion rate of rice based on image recognition according to claim 1, characterized in that: In step S1, the surface crack density of the rice grains is the number of cracks per unit area, and the grain volume change rate is the ratio of the volume of the rice grains at the detection time point to the initial volume of the rice grains. The sum of the normalized surface crack density of the rice grains and the grain volume change rate is used as the damage degree of the rice grains, and the damage degrees of the rice grains in each analysis sample are combined into a damage degree data set.

4. A real-time detection method for the waist explosion rate of rice based on image recognition according to claim 1, characterized in that: In step S2, when generating a label for the detection time interval using the morphological change rate and the damage degree of the rice grains, the least absolute shrinkage and selection operator is used for processing. The specific steps are as follows: Integrate the three input features of each rice image sample into a feature matrix, and the sparsity constraint can be set as the L1 norm penalty term: , and the loss function is: , is the true value of the input feature, is the model output value, n is the number of feature matrices, p is the number of input features in the feature matrix, is the regression coefficient of the model, and λ is the regularization parameter used to control the degree of sparsity.

5. A real-time detection method for the waist explosion rate of rice based on image recognition according to claim 1, characterized in that: In step S4, when the difference between the external damage waist explosion rate of the rice grains and the waist explosion rate caused by the internal stress distribution is within a preset correlation threshold, it is determined that the internal and external factor correlation relationship of the rice grains is strong; otherwise, it is determined that the internal and external factor correlation relationship of the rice grains is weak.

6. A real-time detection method for the waist explosion rate of rice based on image recognition according to claim 5, characterized in that: In step S4, when the internal and external factor correlation relationship of the rice grains in the rice image sample is weak, the random forest algorithm in the decision tree model is used to predict the change trend of the waist explosion rate of the rice grains in the rice image sample; otherwise, the convolutional neural network model under the deep learning framework is used for prediction.

7. A real-time detection system for the waist explosion rate of rice based on image recognition, based on the real-time detection method for the waist explosion rate of rice based on image recognition according to any one of claims 1-6, characterized in that, Including an image acquisition module, a sample analysis module, an environment evaluation module, and a prediction optimization module; The image acquisition module is used to acquire rice image samples on the plane of the sample to be measured; The sample analysis module is used to extract the contour information of the rice grains, remove noise interference, and calculate the area of the waist explosion area; The environment evaluation module is used to acquire the ventilation data of the rice storage environment; The prediction optimization module is used to select different algorithm models according to the internal and external factor correlation relationship and output the waist explosion rate prediction result.

Images