Multi-parameter joint control method and system for key nodes in sea cucumber processing process

Abstract

The present invention provides a multi-parameter joint control method and system for key nodes in the sea cucumber processing process, which relates to the field of automated control technology. The method determines the standard drying characteristics of several drying stages of sea cucumbers; performs drying control of the first drying stage according to preset parameters, monitors and obtains the first sea cucumber image set and extracts the first real-time drying characteristics, and then calculates the drying characteristic deviation; sets the second stage drying target according to the second stage drying index and the drying characteristic deviation, optimizes the drying parameters based on this target, determines the optimal second drying parameters to execute the second drying stage drying control, performs iterative deviation analysis and drying parameter optimization, and completes several drying stages. The present invention solves the technical problem of the prior art that the drying deviation accumulates and the drying uniformity is poor due to the inability to monitor the state of the sea cucumber in real time and dynamically adjust the drying parameters, thereby achieving the technical effect of improving the overall drying uniformity and quality stability of the sea cucumbers.

Description

Technical Field

[0001] The present invention relates to the field of automatic control technology, and in particular to a multi-parameter joint control method and system for key nodes in a sea cucumber processing process. Background Art

[0002] As a high-value seafood, sea cucumber processing, especially the drying process, has a crucial impact on the quality of the final product. Traditional sea cucumber processing methods rely primarily on equipment such as multi-layer belt dryers, which remove moisture through staged heating and air supply. The process uses more traditional control methods, performing drying operations based on preset fixed parameters such as temperature, humidity, and time. This makes it impossible to monitor the actual state changes of the sea cucumber during the drying process in real time, and lacks dynamic feedback on key characteristics such as the sea cucumber's appearance and texture. This makes it difficult to make timely and effective adjustments when deviations occur during the drying process, resulting in accumulated drying deviations, poor drying uniformity, and unstable texture of the finished sea cucumber, affecting the quality of the final product. Summary of the Invention

[0003] The present invention provides a multi-parameter joint control method and system for key nodes in the sea cucumber processing process, which solves the technical problem in the prior art that the sea cucumber status cannot be monitored in real time and the drying parameters cannot be dynamically adjusted during the sea cucumber drying process, resulting in accumulated deviations and poor drying uniformity during the sea cucumber drying process, thereby affecting the sea cucumber drying quality. It achieves the technical effect of realizing efficient closed-loop regulation of the drying process at each stage, thereby improving the overall drying uniformity and quality stability of the sea cucumbers.

[0004] In view of the above problems, on the one hand, the present invention provides a multi-parameter joint control method for key nodes of the sea cucumber processing process, the method comprising: dividing the sea cucumber drying process into several drying stages, and determining several standard drying characteristics; executing the drying control of the first drying stage according to the preset drying parameters, monitoring and obtaining the first sea cucumber image set of the first drying stage; performing drying feature extraction based on the first sea cucumber image set, obtaining the first real-time drying feature, and calculating the first drying feature deviation in combination with the first standard drying feature; setting the second stage drying target according to the second stage drying index and the first drying feature deviation, and optimizing the drying parameters of the second drying stage with the expectation of approaching the second stage drying target, and determining the optimal second drying parameters; executing the sea cucumber drying control of the second drying stage according to the optimal second drying parameters, and performing iterative deviation analysis and drying parameter optimization until all the operations in the several drying stages are completed.

[0005] Preferably, the drying characteristics include size reduction rate, surface grayscale mean and texture density.

[0006] Preferably, drying control of the first drying stage is performed according to preset drying parameters, and the first sea cucumber image set of the first drying stage is monitored and obtained, including: performing drying control of the first drying stage according to preset drying parameters, wherein the drying parameters include at least drying temperature, drying wind speed and relative humidity; at the end of the first drying stage, multi-position image acquisition of the sea cucumbers on the belt dryer is performed by an industrial camera to obtain the first sea cucumber image set.

[0007] Preferably, drying feature extraction is performed based on the first sea cucumber image set to obtain the first real-time drying feature, including: collecting a sample sea cucumber image set based on the historical drying records of sea cucumbers, and performing feature extraction on the sample sea cucumber images according to the drying features to obtain a sample drying feature set; using the sample sea cucumber image set and the sample drying feature set, training a convolutional neural network until convergence to obtain a drying feature identifier; cropping single sea cucumber images from the first sea cucumber image set, selecting complete sea cucumber images to form a first standard single sea cucumber image set; using the drying feature identifier, performing feature extraction on the first standard single sea cucumber image set, and obtaining the first real-time drying feature after mean calculation.

[0008] Preferably, the first sea cucumber image set is cropped with single sea cucumber images, and complete sea cucumber images are selected to form a first standard single sea cucumber image set, including: cropping the first sea cucumber image set with single sea cucumber images, and selecting complete sea cucumber images to form the first single sea cucumber image set; extracting high-frequency difference features from the first single sea cucumber image set to obtain a first standard single sea cucumber image set, wherein, if the similarity between the first single sea cucumber image and other sea cucumber images in the first single sea cucumber image set is less than a preset similarity threshold and the number is greater than a preset number threshold, it is set as the first standard single sea cucumber image.

[0009] Preferably, calculating the first drying characteristic deviation in combination with the first standard drying characteristic includes: obtaining the first standard drying characteristic of the first drying stage; and subtracting the first real-time drying characteristic from the first standard drying characteristic to obtain the first drying characteristic deviation.

[0010] Preferably, the second-stage drying target is set according to the second-stage drying index and the first drying characteristic deviation, and the drying parameters of the second drying stage are optimized with the approach to the second-stage drying target as the expectation, and the optimal second drying parameters are determined, including: obtaining the second standard drying characteristic of the second drying stage, subtracting the first standard drying characteristic of the first drying stage, to obtain the second-stage drying index; obtaining the second-stage drying target by summing the second-stage drying index and the first drying characteristic deviation; based on the drying parameter adjustment space, optimizing the drying parameters of the second drying stage is optimized with the approach to the second-stage drying target as the expectation, and determining the optimal second drying parameters.

[0011] Preferably, based on the drying parameter adjustment space, the drying parameters of the second drying stage are optimized in anticipation of approaching the second stage drying target, and the optimal second drying parameters are determined, including: collecting sample drying parameter sets and sample drying feature sets based on historical sea cucumber drying logs of similar belt dryers, training a generative adversarial network, and constructing a drying feature prediction plug-in; randomly generating a number of drying parameters based on the drying parameter adjustment space, and using the drying feature prediction plug-in to analyze and obtain a number of predicted drying features; taking the second stage drying target as a benchmark, performing an overall deviation analysis on the several predicted drying features to obtain a number of feature deviation values; based on the drying parameter adjustment space, optimizing the drying parameters of the second drying stage according to the several feature deviation values, and outputting the optimal second drying parameters.

[0012] Preferably, based on the drying parameter adjustment space, the drying parameters of the second drying stage are optimized according to the several characteristic deviation values, and the optimal second drying parameters are output, including: taking the drying parameters as the initial solution, sorting them from small to large according to the characteristic deviation values, and mapping to generate several initial solution sequences according to the several characteristic deviation values; dividing the several initial solution sequences into excellent solutions and inferior solutions according to a predetermined ratio, wherein the number of inferior solutions is N times that of the excellent solutions, and N is greater than or equal to 10; clustering the inferior solutions with the excellent solution as the center to obtain multiple solution sets, and in each solution set, taking the excellent solution as the direction, adjusting the inferior solutions in the solution set according to a preset optimization step size to obtain multiple updated solution sets, wherein, if the adjusted inferior solution exceeds the drying parameter adjustment space, a drying parameter is randomly selected from the drying parameter adjustment space for replacement, and if the characteristic deviation value of the adjusted inferior solution is less than the characteristic deviation value of the excellent solution in the same solution set, the inferior solution is used to replace the excellent solution; performing iterative optimization until a preset number of convergences is reached, and outputting the drying parameter with the minimum characteristic deviation value in all solution sets as the optimal second drying parameter.

[0013] On the other hand, the present invention also provides a multi-parameter joint control system for key nodes of the sea cucumber processing process, and the system includes: a drying stage division module, which is used to divide the sea cucumber drying process into several drying stages and determine several standard drying characteristics; a first drying control module, which is used to perform drying control of the first drying stage according to preset drying parameters, and monitor and obtain the first sea cucumber image set of the first drying stage; a feature deviation calculation module, which is used to extract drying features based on the first sea cucumber image set, obtain the first real-time drying feature, and calculate the first drying feature deviation in combination with the first standard drying feature; a drying parameter optimization module, which is used to set the second stage drying target according to the second stage drying index and the first drying feature deviation, and optimize the drying parameters of the second drying stage with the expectation of approaching the second stage drying target, and determine the optimal second drying parameters; a second drying control module, which is used to perform sea cucumber drying control of the second drying stage according to the optimal second drying parameters, and perform iterative deviation analysis and drying parameter optimization until all the operations in the several drying stages are completed.

[0014] One or more technical solutions provided in the present invention have at least the following beneficial effects:

[0015] By dividing the sea cucumber drying process into several drying stages and determining several standard drying characteristics, a stage-by-stage division of the drying process and an ideal drying characteristic benchmark value are provided, providing a reference for subsequent deviation detection and parameter correction. Drying control of the first drying stage is performed according to the preset drying parameters, and the first sea cucumber image set of the first drying stage is monitored and acquired to realize the perception of the sea cucumber state. Drying feature extraction is performed based on the first sea cucumber image set to obtain the first real-time drying feature. The first drying feature deviation is calculated in combination with the first standard drying feature, and the gap between the actual drying result and the standard feature is quantified to clarify the processing deviation. The second-stage drying target is set according to the second-stage drying index and the first drying feature deviation, and the drying parameters of the second drying stage are optimized with the expectation of approaching the second-stage drying target. The optimal second drying parameters are determined, and the actual deviation of the first stage is fed back to the second stage. Dynamic deviation correction is achieved by adjusting the target and parameters of the second stage to ensure that the deviation can be effectively corrected in the subsequent processing stage. The sea cucumber drying control of the second drying stage is executed according to the optimal second drying parameters, and iterative deviation analysis and drying parameter optimization are performed until the several drying stages are completed, forming a multi-stage closed-loop adjustment process: each stage performs parameter optimization and correction based on the deviation of the previous stage, and continuous iteration is performed to avoid deviation accumulation.

[0016] In summary, the present invention divides the sea cucumber drying process into several stages, and obtains the sea cucumber drying characteristics in each stage in combination with real-time visual images, thereby realizing stage-by-stage deviation analysis and dynamic parameter optimization, forming a closed-loop control system of multi-parameter joint control and continuous feedback optimization, significantly improving the uniformity and overall processing quality of the sea cucumber drying process, and realizing efficient, precise and intelligent sea cucumber drying control.

[0017] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flow chart of a multi-parameter joint control method for key nodes in a sea cucumber processing process provided by an embodiment of the present invention.

[0019] Figure 2 A schematic flow chart of obtaining the first real-time drying feature in the multi-parameter joint control method for key nodes in the sea cucumber processing process provided by an embodiment of the present invention.

[0020] Figure 3 A schematic diagram of the structure of a multi-parameter joint control system for key nodes in the sea cucumber processing process provided by an embodiment of the present invention.

[0021] Description of the reference numerals: drying stage division module 10 , first drying control module 20 , characteristic deviation calculation module 30 , drying parameter optimization module 40 , second drying control module 50 . DETAILED DESCRIPTION

[0022] The embodiment of the present invention provides a multi-parameter joint control method and system for key nodes in the sea cucumber processing process, thereby solving the technical problem in the prior art that, due to the inability to monitor the sea cucumber status in real time and dynamically adjust the drying parameters during the sea cucumber drying process, deviations accumulate during the sea cucumber drying process, drying uniformity is poor, and the drying quality of the sea cucumber is affected. The embodiment of the present invention achieves the technical effect of realizing efficient closed-loop regulation of the drying process at each stage, thereby improving the overall drying uniformity and quality stability of the sea cucumber.

[0023] Example 1, as Figure 1 As shown, an embodiment of the present invention provides a multi-parameter joint control method for key nodes in a sea cucumber processing process, the method comprising:

[0024] Step S100: Divide the sea cucumber drying process into several drying stages, and determine several standard drying characteristics.

[0025] Furthermore, the drying characteristics include size reduction rate, surface grayscale mean and texture density.

[0026] Specifically, the sea cucumber drying stage refers to the drying process divided into several continuous or identifiable stages according to factors such as the moisture content, tissue structure changes and appearance characteristics of the sea cucumber material at different time points during the entire sea cucumber drying process. Standard drying characteristics refer to a set of drying characteristic values ​​determined for each drying stage based on historical data or ideal process results, which serves as a target comparison benchmark. Among them, drying characteristics are physical or visual characteristic parameters used to quantitatively evaluate the processing status of sea cucumbers in each drying stage, including three key indicators: size reduction rate, surface grayscale mean and texture density. The size reduction rate indicates the degree of reduction of the sea cucumber in volume or length during the drying process, and is an important indicator reflecting the dehydration rate and texture change; the surface grayscale mean indicates the average brightness of the sea cucumber surface image after graying, reflecting the drying color change and uniformity; the texture density indicates the fineness of the sea cucumber surface texture, which can be obtained through the image texture feature extraction method, reflecting the surface drying quality and product appearance consistency.

[0027] The entire sea cucumber drying process is divided into several drying stages based on factors such as the physical state of the sea cucumber (such as weight and appearance) and the operating characteristics of the drying equipment (such as temperature curve and wind speed curve). In each stage, several standard drying characteristics related to the process quality of that stage are determined, including size reduction rate, surface grayscale mean, and texture density. These standard drying characteristics can be extracted based on historical high-quality drying sample data, and their means can be statistically analyzed to form a benchmark data set. Subsequently, in the process monitoring of each stage, the standard drying characteristics corresponding to each stage are used as the target indicators of drying quality for comparison and correction. For example, the several drying stages divided into the sea cucumber drying process are shown in Table 1:

[0028] Table 1 - Examples of sea cucumber drying stages and standard drying characteristics

[0029]

[0030] By dividing the sea cucumber drying process into stages and determining standard drying characteristics such as size reduction rate, surface grayscale mean and texture density, precise quality target management is achieved for different stages, providing a clear reference and correction direction for the dynamic adjustment of subsequent process parameters, thereby improving drying uniformity and the consistency and stability of the final product.

[0031] Step S200: performing drying control of the first drying stage according to preset drying parameters, and monitoring and acquiring a first sea cucumber image set of the first drying stage.

[0032] Specifically, the preset drying parameters refer to the control variable parameters such as temperature, wind speed and humidity that are set in advance based on historical experience and equipment capabilities during the first stage of drying. The first sea cucumber image set refers to a collection of multiple sea cucumber images collected from multiple positions and angles using an industrial camera after the first drying stage, which is used for subsequent feature extraction and quality assessment. The drying equipment (such as a multi-layer belt dryer) is controlled according to the preset drying parameters to perform the first stage of drying operations. After the first stage is completed, the sea cucumbers are imaged at multiple positions and angles using industrial cameras arranged at different positions of the drying equipment to collect the first sea cucumber image set of the first drying stage. The collected images not only cover the products at different levels of the dryer, but also take into account the multi-perspective characteristics of a single sea cucumber, so as to fully reflect the status characteristics of the sea cucumbers at this stage and ensure the comprehensiveness and accuracy of subsequent feature extraction.

[0033] By collecting images at multiple locations after the first stage, a comprehensive and objective first set of sea cucumber images was obtained, which provided a reliable data basis for subsequent drying feature extraction and real-time deviation analysis, and ensured the scientificity and accuracy of the adjustment and optimization in the next drying stage.

[0034] Step S300: performing drying feature extraction based on the first sea cucumber image set to obtain a first real-time drying feature, and calculating a first drying feature deviation in combination with a first standard drying feature.

[0035] Specifically, the first real-time drying feature refers to the sea cucumber size reduction rate, surface grayscale mean and texture density calculated in real time by an image processing algorithm based on the first sea cucumber image set. The first drying feature deviation refers to the deviation value between the first real-time drying feature and the standard drying feature of the first stage, which is used to quantify the gap between the processing quality of the first stage and the ideal state. First, based on the constructed sample sea cucumber image set and the standard drying feature, a pre-trained convolutional neural network drying feature identifier is used to identify and segment individual sea cucumbers in the first sea cucumber image set, and extract the first real-time drying feature: size reduction rate, surface grayscale mean and texture density. After completing the real-time feature extraction, the deviation value between the real-time feature and the standard feature is calculated in combination with the standard drying feature of the first stage (the first standard drying feature) obtained in step S100, that is, the first real-time drying feature is subtracted from the first standard drying feature to obtain the first drying feature deviation, which provides data support for the target adjustment and parameter optimization in the next stage.

[0036] By analyzing the real-time images of the sea cucumber drying status and extracting the real-time drying features, and comparing them with the standard drying features, the deviation of the sea cucumber drying quality in the first stage can be intuitively quantified, providing a real-time and accurate basis for the target correction and parameter adjustment in the second stage.

[0037] Step S400: setting a second-stage drying target according to the second-stage drying index and the first drying characteristic deviation, and optimizing the drying parameters of the second drying stage with the expectation of approaching the second-stage drying target to determine the optimal second drying parameters.

[0038] Specifically, the second-stage drying index is the initial processing target for the second stage, determined by the difference between the second-stage standard drying characteristic and the first-stage standard drying characteristic. The second-stage drying target is the second-stage drying characteristic value determined based on the second-stage drying index and the deviation from the first-stage drying characteristic, which serves as the target for optimizing the second-stage drying parameters.

[0039] First, based on the difference between the second-stage standard drying characteristics of historical samples and the first-stage standard characteristics, the second-stage drying indicators (e.g., a 30% reduction in size reduction, a 20% reduction in surface grayscale, etc.) are calculated. Next, the deviations from the first-stage drying characteristics are combined to accumulate and calculate the second-stage drying target. Subsequently, based on the adjustable range of drying parameters, a drying feature prediction plug-in trained using a generative adversarial network is used to randomly generate multiple parameter combinations including temperature, wind speed, and humidity within the parameter space and predict the corresponding drying characteristics. Finally, a deviation analysis is performed between the predicted drying characteristics and the second-stage target characteristics. Using an iterative clustering evolutionary optimization algorithm, the optimal second-stage drying parameters are determined with the goal of minimizing the deviation.

[0040] By combining the second-stage indicators with real-time deviations, intelligent analysis and dynamic determination of the optimal second-stage drying parameters are achieved, enabling precise process correction and dynamic process control, ensuring high-quality and stable operation of the second stage.

[0041] Step S500: executing the sea cucumber drying control in the second drying stage according to the optimal second drying parameters, and performing iterative deviation analysis and drying parameter optimization until the several drying stages are completed.

[0042] Specifically, the second drying phase is performed based on the determined optimal second drying parameters. As the second phase progresses, sea cucumber images are continuously collected, real-time drying features are extracted, and deviation analysis is performed to determine the drying target for the next phase and optimize parameters. This process is an iterative closed loop. After each phase, the drying target and parameters for the next phase are dynamically adjusted based on the real-time deviation results until all drying phases are completed, ultimately achieving precise control of the entire process and high-quality product output.

[0043] By combining process control with image analysis, a real-time closed-loop adjustment is formed within and outside the stage, which can suppress the accumulation of deviations in each stage, achieve multi-stage collaborative optimization, and ultimately significantly improve the consistency, uniformity and overall quality of sea cucumber drying.

[0044] Furthermore, step S200 includes:

[0045] Step S210: performing drying control of the first drying stage according to preset drying parameters, wherein the drying parameters at least include drying temperature, drying wind speed and relative humidity.

[0046] Step S220: At the end of the first drying stage, multi-position image acquisition of the sea cucumbers on the belt dryer is performed using an industrial camera to obtain a first sea cucumber image set.

[0047] Specifically, drying temperature refers to the temperature control parameter within the belt dryer, measured in degrees Celsius; drying air speed refers to the airflow velocity within the dryer, measured in m / s; and relative humidity refers to the humidity level within the drying environment, measured in %RH. The belt dryer is controlled according to the preset drying temperature, drying air speed, and relative humidity parameters to begin the first drying phase. At the end of the first drying phase, industrial cameras located inside or outside the dryer capture images of the sea cucumbers from multiple positions (e.g., above, from the side, and diagonally). This captures sea cucumber image data from different perspectives and generates the first sea cucumber image set. This first sea cucumber image set contains a wealth of visual feature information, serving as input for subsequent feature extraction and real-time deviation calculation. For example, in the first drying stage, the drying temperature was set to 55°C, the wind speed was 1.5m / s, and the humidity was 28%RH; after the first stage, three industrial cameras located at the exit of the belt dryer collected images of the dried sea cucumbers from the top, left, and upper right angles, respectively. Each camera collected 50 high-resolution images, resulting in the first sea cucumber image set containing 150 images.

[0048] By executing the above steps, a complete and rich visual dataset of the first drying stage is obtained, providing high-quality input data for subsequent feature extraction and dynamic control, and promoting intelligent and refined management of the sea cucumber drying process.

[0049] Further, such as Figure 2 As shown, step S300 performs drying feature extraction based on the first sea cucumber image set to obtain a first real-time drying feature, including:

[0050] Step S310: collecting a sample sea cucumber image set according to the historical drying records of sea cucumbers, and performing feature extraction on the sample sea cucumber images according to the drying features to obtain a sample drying feature set.

[0051] Step S320: using the sample sea cucumber image set and the sample drying feature set, training a convolutional neural network until convergence to obtain a drying feature identifier.

[0052] Step S330: cropping individual sea cucumber images from the first sea cucumber image set, and selecting complete sea cucumber images to form a first standard single sea cucumber image set.

[0053] Step S340: using the drying feature identifier, extracting features from the first standard single sea cucumber image set, and obtaining a first real-time drying feature after mean calculation.

[0054] Specifically, the sample sea cucumber image set refers to a representative multi-batch, multi-stage sea cucumber image set collected during the historical sea cucumber drying process; the sample drying feature set is a feature data set extracted from the sample sea cucumber image set, including size reduction rate, surface grayscale mean and texture density, etc.; the drying feature identifier is a convolutional neural network model trained based on the sample sea cucumber image set and the sample drying feature set, which is used to automatically extract the drying features of the input sea cucumber image; the first standard single sea cucumber image set is a complete and standardized single sea cucumber image set cropped and screened from the first sea cucumber image set; the first real-time drying feature is extracted and averaged from the first standard single sea cucumber image set, and is used to represent the comprehensive features of the overall drying state of the sea cucumber at the current stage.

[0055] First, a set of sample sea cucumber images was collected from historical drying records, selecting sea cucumber drying processes from different batches and at different temperature and humidity combinations. Each set of sample sea cucumber images included images taken from multiple angles at different locations and time points within the belt dryer. Next, each image was normalized (e.g., to 512×512 pixels) using image processing software, and individual sea cucumber regions were isolated using threshold segmentation, contour detection, or the GrabCut algorithm. For each individual sea cucumber region, the ratio of the image length (measured in pixels) to the initial length (recorded in the database) was calculated to obtain the size reduction ratio. The image was converted to grayscale using the cv2.cvtColor() function, and the pixel mean was obtained using the cv2.mean() function. The sample sea cucumber images were analyzed using the gray-level co-occurrence matrix (GLCM) to generate a co-occurrence matrix, from which texture metrics such as contrast, entropy, or ASM were extracted and finally mapped to texture density values. Drying features extracted from each sample sea cucumber image formed a set of three-dimensional feature vectors, which were summarized as the sample drying feature set.

[0056] The above-mentioned sample sea cucumber image set and its corresponding sample drying feature set constitute a supervised learning dataset. A convolutional neural network is constructed using the TensorFlow or PyTorch framework, with the classic ResNet or VGGNet architecture as the backbone. The input image is 512×512×3 (color image). The network extracts multi-scale features layer by layer through convolution and pooling layers. Finally, a fully connected layer outputs the predicted values ​​for the size reduction ratio, surface grayscale mean, and texture density. The mean squared error (MSE) function is used as the loss function, and the optimizer is Adam or SGD. The learning rate can be initially set to 0.001, and the training batch size is set to 64. The loss function curve is monitored using a validation set. When the loss function value converges and the validation set error reaches the desired threshold (e.g., MSE < 0.01), the model training is considered converged. The model weights are finally saved to form a drying feature identifier.

[0057] For the first sea cucumber image set, image cropping and filtering were used to remove obstructed, blurred, or incomplete images, extracting complete, standard individual sea cucumber images to form the first standard individual sea cucumber image set. Using a trained drying feature identifier, feature extraction was performed on each image in the first standard individual sea cucumber image set. The feature extraction results for all images were then averaged to obtain the mean size reduction rate, surface grayscale mean, and texture density mean of the sea cucumbers in the first stage, forming the first real-time drying feature for the first stage.

[0058] Furthermore, step S330 includes:

[0059] Step S331: cropping individual sea cucumber images from the first sea cucumber image set, and selecting complete sea cucumber images to form a first individual sea cucumber image set.

[0060] Step S332: extract high-frequency difference features from the first single sea cucumber image set to obtain a first standard single sea cucumber image set, wherein if the similarity between the first single sea cucumber image and other sea cucumber images in the first single sea cucumber image set is less than a preset similarity threshold and the number is greater than a preset number threshold, it is set as the first standard single sea cucumber image.

[0061] Specifically, a complete sea cucumber image refers to an image in which the sea cucumber has a complete shape, is not blocked by other objects, has no damaged or missing parts, and has high image clarity, which can accurately reflect the appearance characteristics of the sea cucumber. High-frequency difference feature extraction refers to extracting features from the image that have a high frequency of change and can reflect significant differences between individuals, such as texture details, edge features, etc. The preset similarity threshold is a pre-set similarity threshold used to determine whether two images have similar features. The preset number threshold is a pre-set number threshold used to screen typical feature images with significant differences.

[0062] The contour detection algorithm in the OpenCV library is used to detect the boundaries of the sea cucumber in the image and generate a bounding rectangle or polygonal region. Furthermore, the bounding box coordinate information of a single sea cucumber is obtained using cv2.boundingRect() or cv2.minAreaRect(). The sea cucumber region is cropped according to the bounding box to remove the background. After cropping, the individual sea cucumber images are uniformly resized (for example, standardized to 512×512 pixels) and stored in the first single sea cucumber image set. This cropping and standardization ensures that each image contains only a single complete sea cucumber and preserves the surface morphology and texture features of the sea cucumber, facilitating subsequent feature extraction.

[0063] For each image in the first individual sea cucumber image set, a two-dimensional fast Fourier transform (FFT) is used to extract high-frequency features. Specifically, a Fourier transform is performed on each individual sea cucumber image using numpy.fft.fft2() or cv2.dft() to obtain a frequency domain representation. Furthermore, the spectrum is centered using the numpy.fft.fftshift() function, concentrating the high-frequency components at the edges of the spectrum. Next, a high-frequency component feature vector is extracted. Preferably, the modulus and phase distribution of the portion exceeding a set spectral radius threshold are statistically analyzed to form a high-frequency feature vector. Next, the similarity between each individual sea cucumber image and other images in the first individual sea cucumber image set is calculated using cosine similarity or Euclidean distance. Preferably, a similarity threshold (e.g., 0.8) is set. If the number of times a single sea cucumber image's similarity to other images is less than this threshold exceeds a preset threshold (e.g., 3), the image is identified as a representative feature image with significant differences from the overall image set. Through these steps, several representative individual sea cucumber images with significant high-frequency feature differences are selected to form the first standard individual sea cucumber image set. By analyzing and screening high-frequency features, the surface features of sea cucumbers in the typical drying stage can be effectively extracted, reducing redundant calculations in full data processing and improving the accuracy and efficiency of feature extraction.

[0064] Furthermore, in step S300, the first drying characteristic deviation is calculated in combination with the first standard drying characteristic, including:

[0065] Step S350: obtaining a first standard drying characteristic of the first drying stage.

[0066] Step S360: subtracting the first real-time drying characteristic from the first standard drying characteristic to obtain a first drying characteristic deviation.

[0067] Specifically, the first standard drying feature is a drying feature reference value formed based on the typical surface features of the sea cucumber under the standard process conditions of the first drying stage. Preferably, the first standard drying feature can be obtained by the following method: through previous multiple batches of drying experiments, the typical process parameters of each stage and their corresponding sea cucumber image features are recorded, and combined with the quality inspection results of the sea cucumber (such as moisture content, appearance, elasticity, etc.), the drying image and its extracted features in the optimal state in the first drying stage are screened out as the first standard drying feature. In the specific implementation process, based on three-dimensional feature parameters such as sea cucumber surface texture density, size reduction rate and surface grayscale mean, the mean or median of multiple experimental results can be taken to form a multidimensional vector of the first standard drying feature. Through the above-mentioned acquisition method, it is ensured that the first standard drying feature can fully reflect the ideal drying state of the sea cucumber in the first drying stage, serving as a benchmark for subsequent feature deviation analysis.

[0068] The first real-time drying feature vector extracted by the drying feature identifier is subjected to a difference operation with the first standard drying feature vector obtained. Preferably, an element-by-element difference method is adopted. The specific formula is as follows: first drying feature deviation = first standard drying feature - first real-time drying feature. Among them, the three-dimensional vector dimensions correspond to texture density, size reduction rate and surface grayscale mean, respectively. The three-dimensional difference result can directly represent the degree of deviation of each feature dimension. The numerical unit of the difference result is the same as that of the original feature. By calculating the first drying feature deviation, the difference between the current drying state and the ideal state can be quantified, providing a basis for subsequent deviation feedback and parameter optimization, and realizing dynamic adjustment and refined control of the drying process.

[0069] Furthermore, step S400 includes:

[0070] Step S410: Obtain the second standard drying characteristic of the second drying stage, subtract the first standard drying characteristic of the first drying stage, and obtain the second stage drying index.

[0071] Step S420: obtaining a second-stage drying target according to the sum of the second-stage drying index and the first drying characteristic deviation.

[0072] Step S430: Based on the drying parameter adjustment space, optimizing the drying parameters for the second drying stage is performed in order to approach the second drying target, and determining the optimal second drying parameters.

[0073] Specifically, the second standard drying feature is the optimal drying feature vector (a three-dimensional vector consisting of texture density, size reduction ratio, and surface grayscale mean) for the second drying stage, similarly based on multi-batch experimental results and quality inspection data. The second standard drying feature is obtained in the same manner as the first standard drying feature. The second standard drying feature of the second drying stage is subtracted from the first standard drying feature of the first drying stage to obtain the second-stage drying index, which visually reflects the ideal direction and magnitude of change during the transition from the first stage to the second stage.

[0074] To ensure that the characteristic evolution of the second drying stage is as close to the ideal transition effect as possible, the second-stage drying index and the first-stage drying characteristic deviation are added element-by-element to form the second-stage drying target vector. The formula is: Second-stage Drying Target = Second-stage Drying Index + First-stage Drying Characteristic Deviation. This second-stage drying target comprehensively reflects the deviation between the current drying state and the ideal stage and provides a dynamic target value for the next step of drying parameter optimization.

[0075] The drying parameter adjustment space refers to the range within which drying parameters (drying temperature, drying air speed, and relative humidity) can vary within the equipment's permitted range. This provides a feasible range for optimizing drying parameters. Based on the drying equipment's performance and drying process specifications, the adjustable ranges for drying temperature, drying air speed, and relative humidity are determined to construct the drying parameter adjustment space. For example, the temperature range is 40°C to 60°C, the air speed range is 1m / s to 3m / s, and the relative humidity range is 50% to 70% RH. Based on this, a goal-driven optimization algorithm (such as gradient descent or genetic algorithm) is used to iteratively update the parameters, simulating the predicted feature evolution trajectory until the Euclidean distance or weighted deviation between the predicted drying feature vector and the second-stage drying target vector is minimized, or a preset deviation threshold (e.g., less than 1%) is met. Ultimately, the optimal second-stage drying parameters are output as the actual execution parameters for the second drying stage, enabling intelligent and adaptive dynamic adjustment of drying parameters.

[0076] Furthermore, step S430 includes:

[0077] Step S431: Based on the historical sea cucumber drying logs of similar belt dryers, a sample drying parameter set and a sample drying feature set are collected, a generative adversarial network is trained, and a drying feature prediction plug-in is constructed.

[0078] Step S432: randomly generating a number of drying parameters based on the drying parameter adjustment space, and using the drying feature prediction plug-in to analyze and obtain a number of predicted drying features.

[0079] Step S433: Based on the second-stage drying target, an overall deviation analysis is performed on the plurality of predicted drying features to obtain a plurality of feature deviation values.

[0080] Step S434: Based on the drying parameter adjustment space, optimizing the drying parameters of the second drying stage according to the plurality of characteristic deviation values, and outputting the optimal second drying parameters.

[0081] Specifically, based on historical sea cucumber drying logs from similar belt dryers, we collected drying parameter sets (including temperature, wind speed, humidity, etc.) and corresponding drying feature sets (such as size reduction rate, surface grayscale mean, and texture density) for multiple batches of samples under different drying parameters. Using a generative adversarial network for training and optimizing the generator-discriminator game, we constructed a drying feature prediction plug-in with strong generalization capabilities. This plug-in rapidly generates corresponding drying feature vector predictions for given drying parameters, enabling rapid estimation of drying process feature evolution under different parameter configurations.

[0082] Within the drying parameter adjustment space, several drying parameter combinations are randomly generated. For each generated parameter set, the drying feature prediction plug-in is called to obtain the corresponding predicted drying feature. The predicted result is also a three-dimensional vector, which can be directly compared and analyzed with the second-stage drying target.

[0083] Using the second-stage drying target as a benchmark, the Euclidean distance or weighted absolute deviation between each predicted drying feature and the target vector is calculated to obtain several feature deviation values. The smaller the deviation value, the closer the predicted feature is to the second-stage drying target. Based on these feature deviation values, the drying parameters are optimized within the drying parameter adjustment space. The parameter combination with the smallest deviation value is selected and output as the optimal second-stage drying parameters.

[0084] Furthermore, step S434 includes:

[0085] Step S434 - 1 : Taking the drying parameters as the initial solution, sorting the solution from small to large according to the characteristic deviation values, and mapping and generating a plurality of initial solution sequences according to the plurality of characteristic deviation values.

[0086] Step S434-2: Divide the plurality of initial solution sequences into superior solutions and inferior solutions according to a predetermined ratio, wherein the number of inferior solutions is N times the number of superior solutions, where N is greater than or equal to 10.

[0087] Step S434-3: Cluster the inferior solutions with the optimal solution as the center to obtain multiple solution sets, and within each solution set, adjust the inferior solutions in the solution set with the optimal solution as the direction according to the preset optimal step length to obtain multiple updated solution sets, wherein if the adjusted inferior solution exceeds the drying parameter adjustment space, a drying parameter is randomly selected in the drying parameter adjustment space for replacement, and if the characteristic deviation value of the adjusted inferior solution is less than the characteristic deviation value of the optimal solution in the same solution set, the inferior solution is used to replace the optimal solution.

[0088] Step S434-4: perform iterative optimization until a preset number of convergences is reached, and output the drying parameter with the minimum characteristic deviation value in all solution sets as the optimal second drying parameter.

[0089] Specifically, the drying parameters and their corresponding characteristic deviation values ​​obtained in step S433 are sorted from smallest to largest according to their characteristic deviation values. Based on the sorting results, several initial solution sequences are mapped and generated. Each initial solution sequence represents an initial drying parameter combination, serving as a starting point for subsequent iterative optimization. The initial solution sequences are divided into a set of superior solutions and a set of inferior solutions according to a predetermined ratio. The superior solution set is a subset of initial solutions with relatively low characteristic deviation values. The inferior solution set is the set of initial solutions remaining after removing the superior solution set from the initial solution sequences. The number of such solutions is N times the number of the superior solution sets, where N ≥ 10 (e.g., 10 superior solutions and 100 inferior solutions).

[0090] With each optimal solution as the center, the inferior solutions are clustered to form multiple solution sets. Within each solution set, adjustments are made according to the following steps: Focusing on the optimal solution, the inferior solutions within the solution set are fine-tuned according to a preset optimization step size (e.g., a normalized ratio of 0.01 to 0.1) to generate an updated solution set. If the adjusted inferior solution exceeds the drying parameter adjustment space (i.e., the physically or technologically acceptable temperature, wind speed, and humidity range), a new set of feasible drying parameters is randomly selected within that drying parameter adjustment space to replace it and ensure the physical feasibility of the solution. The characteristic deviation values ​​of the updated inferior solution are compared with those of the optimal solution within the same solution set. If the inferior solution outperforms the original optimal solution (i.e., has a smaller characteristic deviation value), it is replaced with the current optimal solution. Through continuous adjustments and optimal solution updates, each iteration ensures that the solution space with smaller characteristic deviation values ​​is reached.

[0091] Repeat the above adjustment-replacement process and perform multiple rounds of iterative optimization until the preset number of convergences is reached. Finally, among all solution sets, the drying parameter corresponding to the minimum characteristic deviation value is extracted as the optimal second drying parameter for the second drying stage.

[0092] In summary, the multi-parameter joint control method for key nodes in the sea cucumber processing process provided by the embodiment of the present invention has the following beneficial effects:

[0093] By dividing the sea cucumber drying process into several drying stages and determining several standard drying characteristics, a stage-by-stage division of the drying process and an ideal drying characteristic benchmark value are provided, providing a reference for subsequent deviation detection and parameter correction. Drying control of the first drying stage is performed according to the preset drying parameters, and the first sea cucumber image set of the first drying stage is monitored and acquired to realize the perception of the sea cucumber state. Drying feature extraction is performed based on the first sea cucumber image set to obtain the first real-time drying feature. The first drying feature deviation is calculated in combination with the first standard drying feature, and the gap between the actual drying result and the standard feature is quantified to clarify the processing deviation. The second-stage drying target is set according to the second-stage drying index and the first drying feature deviation, and the drying parameters of the second drying stage are optimized with the expectation of approaching the second-stage drying target. The optimal second drying parameters are determined, and the actual deviation of the first stage is fed back to the second stage. Dynamic deviation correction is achieved by adjusting the target and parameters of the second stage to ensure that the deviation can be effectively corrected in the subsequent processing stage. The sea cucumber drying control of the second drying stage is executed according to the optimal second drying parameters, and iterative deviation analysis and drying parameter optimization are performed until the several drying stages are completed, forming a multi-stage closed-loop adjustment process: each stage performs parameter optimization and correction based on the deviation of the previous stage, and continuous iteration is performed to avoid deviation accumulation.

[0094] Overall, the embodiment of the present invention achieves stage-by-stage deviation by dividing the sea cucumber drying process into several stages and combining real-time visual images to obtain sea cucumber drying characteristics at each stage. At the same time, the adaptive iterative optimization based on multi-solution clustering effectively overcomes the local optimal trap, realizes dynamic adjustment of optimal parameters and global quality balance in each stage of the drying process, and forms a closed-loop control system of multi-parameter joint control and continuous feedback optimization, which significantly improves the uniformity and overall processing quality of the sea cucumber drying process and realizes efficient, precise and intelligent sea cucumber drying control.

[0095] Example 2, as Figure 3 As shown, based on the same inventive concept as the aforementioned embodiment 1, this embodiment of the present invention provides a multi-parameter joint control system for key nodes in the sea cucumber processing process, the system comprising:

[0096] The drying stage division module 10 is used to divide the sea cucumber drying process into several drying stages and determine several standard drying characteristics.

[0097] The first drying control module 20 is used to perform drying control in the first drying stage according to preset drying parameters, and monitor and obtain a first sea cucumber image set in the first drying stage.

[0098] The feature deviation calculation module 30 is used to extract drying features based on the first sea cucumber image set, obtain a first real-time drying feature, and calculate a first drying feature deviation in combination with the first standard drying feature.

[0099] The drying parameter optimization module 40 is used to set the second stage drying target according to the second stage drying index and the first drying characteristic deviation, and optimize the drying parameters of the second drying stage with the expectation of approaching the second stage drying target to determine the optimal second drying parameters.

[0100] The second drying control module 50 is used to execute the sea cucumber drying control in the second drying stage according to the optimal second drying parameters, and perform iterative deviation analysis and drying parameter optimization until the several drying stages are completed.

[0101] Furthermore, the drying characteristics include size reduction rate, surface grayscale mean and texture density.

[0102] Furthermore, the first drying control module 20 of the embodiment of the present invention is further configured to perform the following steps:

[0103] Drying control of the first drying stage is performed according to preset drying parameters, wherein the drying parameters include at least drying temperature, drying wind speed and relative humidity; at the end of the first drying stage, multi-position image acquisition of the sea cucumber on the belt dryer is performed by an industrial camera to obtain a first sea cucumber image set.

[0104] Furthermore, the characteristic deviation calculation module 30 of the embodiment of the present invention is further configured to perform the following steps:

[0105] According to the historical drying records of sea cucumbers, a sample sea cucumber image set is collected, and feature extraction is performed on the sample sea cucumber images according to the drying features to obtain a sample drying feature set; the sample sea cucumber image set and the sample drying feature set are used to train a convolutional neural network until convergence to obtain a drying feature identifier; single sea cucumber images are cropped from the first sea cucumber image set, and complete sea cucumber images are selected to form a first standard single sea cucumber image set; using the drying feature identifier, feature extraction is performed on the first standard single sea cucumber image set, and a first real-time drying feature is obtained after mean calculation.

[0106] Furthermore, the characteristic deviation calculation module 30 of the embodiment of the present invention is further configured to perform the following steps:

[0107] The first sea cucumber image set is cropped for individual sea cucumber images, and complete sea cucumber images are selected to form a first individual sea cucumber image set; high-frequency difference features are extracted for the first individual sea cucumber image set to obtain a first standard individual sea cucumber image set, wherein if the number of similarities between the first individual sea cucumber image and other sea cucumber images in the first individual sea cucumber image set is less than a preset similarity threshold and is greater than a preset number threshold, then the first standard individual sea cucumber image is set.

[0108] Furthermore, the characteristic deviation calculation module 30 of the embodiment of the present invention is further configured to perform the following steps:

[0109] A first standard drying characteristic of a first drying stage is obtained; and the first real-time drying characteristic is subtracted from the first standard drying characteristic to obtain a first drying characteristic deviation.

[0110] Furthermore, the drying parameter optimization module 40 of the embodiment of the present invention is further configured to perform the following steps:

[0111] Obtain a second standard drying characteristic for the second drying stage, subtract the first standard drying characteristic for the first drying stage, and obtain a second-stage drying index; obtain a second-stage drying target based on the sum of the deviations between the second-stage drying index and the first drying characteristic; and optimize the drying parameters for the second drying stage based on the drying parameter adjustment space, with the aim of approaching the second-stage drying target, to determine the optimal second drying parameters.

[0112] Furthermore, the drying parameter optimization module 40 of the embodiment of the present invention is further configured to perform the following steps:

[0113] Based on the historical sea cucumber drying logs of similar belt dryers, sample drying parameter sets and sample drying feature sets are collected, a generative adversarial network is trained, and a drying feature prediction plug-in is constructed; based on the drying parameter adjustment space, several drying parameters are randomly generated, and the drying feature prediction plug-in is used to analyze and obtain several predicted drying features; based on the second-stage drying target, an overall deviation analysis is performed on the several predicted drying features to obtain several feature deviation values; based on the drying parameter adjustment space, the drying parameters of the second drying stage are optimized according to the several feature deviation values, and the optimal second drying parameters are output.

[0114] Furthermore, the drying parameter optimization module 40 of the embodiment of the present invention is further configured to perform the following steps:

[0115] Taking the drying parameter as the initial solution, sorting them from small to large according to the characteristic deviation value, and mapping to generate several initial solution sequences according to the several characteristic deviation values; dividing the several initial solution sequences into excellent solutions and inferior solutions according to a predetermined ratio, wherein the number of inferior solutions is N times that of the excellent solutions, and N is greater than or equal to 10; clustering the inferior solutions with the excellent solution as the center to obtain multiple solution sets, and in each solution set, taking the excellent solution as the direction, adjusting the inferior solutions in the solution set according to the preset optimization step size to obtain multiple updated solution sets, wherein, if the adjusted inferior solution exceeds the drying parameter adjustment space, a drying parameter is randomly selected in the drying parameter adjustment space for replacement, and if the characteristic deviation value of the adjusted inferior solution is less than the characteristic deviation value of the excellent solution in the same solution set, the inferior solution is used to replace the excellent solution; performing iterative optimization until the preset number of convergences is reached, and outputting the drying parameter with the minimum characteristic deviation value in all solution sets as the optimal second drying parameter.

[0116] Through the detailed description of the multi-parameter joint control method for key nodes in the sea cucumber processing process mentioned above in this specification, those skilled in the art can clearly understand the multi-parameter joint control system for key nodes in the sea cucumber processing process in this embodiment. For the system disclosed in Example 2, since it corresponds to the method disclosed in Example 1 and has corresponding functional modules and beneficial effects, the relevant details can be referred to the description of the method part.

[0117] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-parameter joint control method for key nodes in the sea cucumber processing process, characterized in that: Methods include: The sea cucumber drying process is divided into several drying stages, and several standard drying characteristics are determined; Executing drying control of a first drying stage according to preset drying parameters, and monitoring and acquiring a first sea cucumber image set of the first drying stage; Extracting a drying feature based on the first sea cucumber image set to obtain a first real-time drying feature, and calculating a first drying feature deviation based on the first standard drying feature; Obtaining a second standard drying characteristic for the second drying stage, subtracting the first standard drying characteristic for the first drying stage to obtain a second-stage drying index, setting a second-stage drying target based on a deviation between the second-stage drying index and the first drying characteristic, and optimizing drying parameters for the second drying stage with the expectation of approaching the second-stage drying target to determine optimal second drying parameters; Execute the sea cucumber drying control of the second drying stage according to the described optimal second drying parameter, and perform iterative deviation analysis and drying parameter optimization until the described several drying stages are all completed; The method further comprises: setting a second-stage drying target according to the second-stage drying index and the first drying characteristic deviation, optimizing drying parameters for the second drying stage with the expectation of approaching the second-stage drying target, and determining optimal second drying parameters, including: Obtaining a second-stage drying target based on the sum of the second-stage drying index and the first drying characteristic deviation; Based on the drying parameter adjustment space, optimizing the drying parameters for the second drying stage in anticipation of approaching the second drying target, and determining the optimal second drying parameters; Determining the optimal second drying parameter includes: Based on the historical sea cucumber drying logs of similar belt dryers, we collected sample drying parameter sets and sample drying feature sets, trained a generative adversarial network, and built a drying feature prediction plug-in. Randomly generating a number of drying parameters based on the drying parameter adjustment space, and obtaining a number of predicted drying characteristics by analyzing the drying characteristic prediction plug-in; Based on the second-stage drying target, an overall deviation analysis is performed on the plurality of predicted drying characteristics to obtain a plurality of characteristic deviation values; Based on the drying parameter adjustment space, the drying parameters of the second drying stage are optimized according to the plurality of characteristic deviation values, and the optimal second drying parameters are output.

2. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 1 is characterized in that: The drying characteristics include size reduction rate, surface grayscale mean and texture density.

3. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 1 is characterized in that: The method includes performing drying control of the first drying stage according to preset drying parameters, and monitoring and obtaining a first sea cucumber image set of the first drying stage, including: Executing drying control of the first drying stage according to preset drying parameters, wherein the drying parameters include at least drying temperature, drying wind speed and relative humidity; At the end of the first drying stage, multi-position images of the sea cucumbers on the belt dryer are collected by an industrial camera to obtain a first sea cucumber image set.

4. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 2 is characterized in that: Extracting drying features according to the first sea cucumber image set to obtain first real-time drying features includes: According to the historical drying records of sea cucumbers, a sample sea cucumber image set is collected, and feature extraction is performed on the sample sea cucumber images according to the drying features to obtain a sample drying feature set; Using the sample sea cucumber image set and the sample drying feature set, a convolutional neural network is trained until convergence to obtain a drying feature identifier; performing single sea cucumber image cropping on the first sea cucumber image set, and selecting complete sea cucumber images to form a first standard single sea cucumber image set; The drying feature identifier is used to perform feature extraction on the first standard single sea cucumber image set, and the first real-time drying feature is obtained after mean calculation.

5. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 4 is characterized in that: Cropping individual sea cucumber images from the first sea cucumber image set, and selecting complete sea cucumber images to form a first standard individual sea cucumber image set, comprising: performing single sea cucumber image cropping on the first sea cucumber image set, and selecting complete sea cucumber images to form a first single sea cucumber image set; High-frequency difference features are extracted from the first single sea cucumber image set to obtain a first standard single sea cucumber image set, wherein if the similarity between the first single sea cucumber image and other sea cucumber images in the first single sea cucumber image set is less than a preset similarity threshold and the number is greater than a preset number threshold, it is set as the first standard single sea cucumber image.

6. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 1 is characterized in that: The first drying characteristic deviation is calculated in combination with the first standard drying characteristic, including: obtaining a first standard drying characteristic of a first drying stage; The first real-time drying characteristic is subtracted from the first standard drying characteristic to obtain a first drying characteristic deviation.

7. The multi-parameter joint control method for key nodes in the sea cucumber processing process according to claim 1 is characterized in that: Based on the drying parameter adjustment space, optimizing the drying parameters of the second drying stage according to the plurality of characteristic deviation values, and outputting the optimal second drying parameters, including: Taking the drying parameters as the initial solution, sorting them from small to large according to the characteristic deviation values, and mapping and generating several initial solution sequences according to the several characteristic deviation values; Dividing the plurality of initial solution sequences into superior solutions and inferior solutions according to a predetermined ratio, wherein the number of inferior solutions is N times the number of superior solutions, where N is greater than or equal to 10; Clustering the inferior solutions with the optimal solution as the center to obtain multiple solution sets. Within each solution set, adjusting the inferior solutions in the solution set according to a preset optimization step length with the optimal solution as the direction to obtain multiple updated solution sets. If the adjusted inferior solution exceeds the drying parameter adjustment space, a drying parameter is randomly selected from the drying parameter adjustment space for replacement. If the characteristic deviation value of the adjusted inferior solution is less than the characteristic deviation value of the optimal solution in the same solution set, the superior solution is replaced by the inferior solution. Iterate the optimization until the preset number of convergences is reached, and output the drying parameter with the minimum characteristic deviation value in all solution sets as the optimal second drying parameter.

8. A multi-parameter joint control system for key nodes in the sea cucumber processing process, characterized by: The system is used to execute the multi-parameter joint control method for key nodes in the sea cucumber processing process according to any one of claims 1 to 7, comprising: A drying stage division module is used to divide the sea cucumber drying process into several drying stages and determine several standard drying characteristics; A first drying control module is used to perform drying control in a first drying stage according to preset drying parameters, and monitor and obtain a first sea cucumber image set in the first drying stage; a feature deviation calculation module, configured to extract a drying feature based on the first sea cucumber image set, obtain a first real-time drying feature, and calculate a first drying feature deviation based on the first standard drying feature; a drying parameter optimization module, configured to obtain a second standard drying characteristic for the second drying stage, subtract the first standard drying characteristic for the first drying stage from the second standard drying characteristic to obtain a second-stage drying index, set a second-stage drying target based on a deviation between the second-stage drying index and the first drying characteristic, and optimize the drying parameters for the second drying stage with the expectation of approaching the second-stage drying target to determine the optimal second drying parameters; The second drying control module is used to execute the sea cucumber drying control in the second drying stage according to the optimal second drying parameters, and perform iterative deviation analysis and drying parameter optimization until the operations in the several drying stages are completed.

Images