Intelligent broccoli storage and fresh-keeping environment control system

Abstract

The invention relates to the technical field of vegetable preservation, and discloses a broccoli intelligent storage and preservation environment control system, which comprises an intelligent management system, and is characterized in that the intelligent management system comprises a storage environment control module, an online multi-mode quality detection module, an analysis feedback module, an intelligent decision execution unit and an early warning reminding module; the online multi-mode quality detection module, the analysis feedback module, the intelligent decision execution unit and the execution mechanism are integrated to construct a set of complete fresh-keeping feedback control ring, and meanwhile, key indexes such as mildew, water loss rate, color, ethylene concentration and the like can be subjected to automatic online full detection; differentiated fresh-keeping measures such as precise humidity and temperature adjustment, modified atmosphere packaging and isolation treatment are automatically triggered based on the judgment result, the conventional extensive management mode depending on manual spot check and passive response is changed, problems can be found in real time and actively intervened in the storage process, loss is reduced, and the shelf life is prolonged.

Description

Technical Field

[0001] This invention relates to the field of vegetable preservation technology, and more specifically to an intelligent storage and preservation environment control system for broccoli. Background Technology

[0002] Broccoli, as a highly nutritious vegetable, undergoes vigorous respiration and metabolism after harvest, making it extremely susceptible to water loss, wilting, yellowing, mold, and spoilage. Its preservation quality directly impacts its commercial value, food safety, and supply chain efficiency. With the increasing demand for high-quality fresh vegetables in the consumer market, higher requirements are being placed on the precise and intelligent management of broccoli post-harvest preservation. However, traditional broccoli storage and preservation methods still have the following prominent limitations: Quality monitoring methods are limited and outdated: key deterioration indicators (such as internal mold, water loss rate, color change, and ethylene accumulation) mostly rely on manual visual inspection or offline sampling, which is inefficient and has a narrow coverage. It is impossible to achieve real-time and full-area monitoring within the warehouse, and it is difficult to provide early warning for hidden problems such as internal mold. It also lacks quantitative tracking of continuous changes such as water loss and yellowing. Environmental control is crude and slow to respond: the warehouse environment mostly adopts fixed settings or simple time-series control, which cannot be dynamically and accurately controlled according to the real-time physiological state and quality changes of broccoli. When local wilting or ethylene accumulation is found, the entire warehouse area can only be adjusted uniformly, which is energy-intensive and lacks specificity, and cannot achieve rapid intervention and rescue for problematic batches. Therefore, there is an urgent need for an intelligent broccoli storage and preservation environment control system to solve the aforementioned technical problems. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a broccoli intelligent storage and preservation environment control system to solve the technical problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a broccoli intelligent storage and preservation environment control system, including an intelligent management system, wherein the intelligent management system includes a storage environment control module, an online multimodal quality detection module, an analysis and feedback module, an intelligent decision execution unit, and an early warning and reminder module; The storage environment control module is used to automatically control the environmental parameters of temperature, humidity, and gas composition in the broccoli storage warehouse and to set target parameters; The online multimodal quality inspection module is set up in the receiving area or between storage shelves in the warehouse to perform real-time, non-destructive inspection on broccoli in storage or circulation in order to collect its multidimensional quality data. The analysis and feedback module is connected to the online multimodal quality detection module. It is used to receive and analyze multidimensional quality data and generate information feedback on the broccoli quality and safety status assessment results. The intelligent decision execution unit is communicatively connected to the analysis feedback module and the warehouse environment control module. It is used to generate corresponding control decisions based on the information feedback and transmit the control decisions to the warehouse environment control module. The warehouse environment control module drives the execution mechanism set inside it to perform environmental regulation or physical operation according to the control decisions, so as to intervene or isolate the broccoli batches with abnormal quality or safety risks. The early warning module is connected to the intelligent decision execution unit and the analysis feedback module, and is used to send the information feedback and the control decision to a remote terminal.

[0005] Preferably, the quality abnormalities include: mold, excessive water loss, abnormal browning, and abnormal ethylene concentration; Preferably, the warehouse environment control module includes an environment control sub-module set in each area of ​​the broccoli warehouse, which is used to control the environment of each area to operate according to the set target parameters; The system also includes a main storage flow channel, wherein the actuator is connected in parallel with or bypasses the main storage flow channel, and the actuator includes: The precise humidity and temperature control submodule is used to perform localized humidification and precise cooling treatment on batches of broccoli that have excessive water loss or are at risk of wilting. The modified atmosphere packaging and gas control submodule is used to automatically adjust the gas composition inside the packaging or in a local space based on ethylene concentration and respiration intensity data, so as to carry out intelligent modified atmosphere preservation. The color and mold intelligent sorting submodule is used to integrate hyperspectral and image data to identify areas of abnormal browning or mold, and to intelligently sort and isolate broccoli. The intelligent isolation and recycling submodule is used to automatically transfer, isolate, and dispose of abnormal batches of broccoli that show clear signs of mold or severe spoilage.

[0006] Preferably, the intelligent decision-making execution unit is further configured to: plan an ordered processing path for each batch of abnormalities in the execution mechanism based on the quality anomaly type provided by the analysis feedback module, and control each sub-module of the execution mechanism to run sequentially according to this path. The planning of the processing path follows these steps: S11. If mold is detected, regardless of whether other abnormalities exist at the same time, the batch will be directly routed to the intelligent isolation and recycling submodule. S12. If no mold is detected and the water loss rate exceeds the standard, the system will be routed to the precision humidity and temperature regulation submodule for priority processing, and a decision will be made based on the processing result as to whether to enter the subsequent submodule or return to the main storage area. S13. If no mold growth or excessive water loss is detected, but there is abnormal browning or abnormal ethylene concentration, then the device is planned to be routed to the modified atmosphere packaging and gas control submodule. S14. If only local color abnormalities are detected and there is no mold, then the sample is routed to the intelligent color and mold sorting submodule for sorting, and the qualified sample is returned to storage. All broccoli batches processed through the planning paths S12 to S14 are ultimately routed to the appropriate storage area that meets their quality requirements for subsequent preservation.

[0007] Preferably, the online multimodal quality detection module integrates multiple sub-units for collecting multidimensional quality data of broccoli, wherein the multiple sub-units include: The near-infrared spectroscopy detection subunit emits near-infrared light in a specific band and receives the diffuse reflectance spectrum of broccoli. It calculates and outputs the water loss rate value Wr of broccoli in real time through the built-in spectral analysis model. The hyperspectral imaging subunit scans the surface of broccoli online to acquire high-dimensional image cube data containing continuous spectral information. Through spectral feature matching and classification algorithms, it simultaneously extracts the feature index Mr, the color browning index Br, and the relative chlorophyll content Cr from the data to characterize the risk of mold. The machine vision image acquisition subunit acquires high-definition color images of broccoli under uniform lighting conditions. It analyzes the compactness of the florets, color uniformity, and surface defects through image processing algorithms, and then calculates the morphological score Sr and the proportion of local browning area Lr. The gas sensing subunit monitors the ethylene concentration Er, carbon dioxide concentration CO2r, and relative humidity RHr in the broccoli storage microenvironment.

[0008] Preferably, the near-infrared spectroscopy detection subunit, hyperspectral imaging subunit, machine vision image acquisition subunit, and gas sensing subunit are distributed and deployed at key nodes in the storage area, including inbound inspection stations, storage shelf areas, and outbound verification points, for detecting broccoli in static storage or during turnover. Each subunit is started by the system according to the detection task, and a unified data association mechanism ensures that the multi-dimensional detection data of the same batch of broccoli is consistent in time and space.

[0009] Preferably, the analysis feedback module is configured to perform the following analysis steps to process the multidimensional quality data collected from the online multimodal quality detection module, the specific process of which is as follows: S100, Data Reception: Receive and associate multidimensional quality data from the online multimodal quality detection module, the data including at least water loss rate Wr, mold characteristic index Mr, browning index Br, chlorophyll content Cr, morphology score Sr, browning area percentage Lr, and ethylene concentration Er. S200, Threshold Comparison: Compare the received quantitative data with the corresponding quality thresholds preset in the database, including: comparing Wr with the water loss threshold Ws, comparing Mr with the mold threshold Ms, comparing Br with the browning threshold Bs, comparing Sr with the morphology threshold Ss, comparing Lr with the local browning threshold Ls, and comparing Er with the ethylene threshold Es. S300: Execute intelligent decision based on the comparison results: If Mr > Ms, it is determined to be a safety anomaly, and the first type of anomaly information feedback is generated, which includes the mold characteristic index Mr and its exceeding data; If Wr > Ws, it is determined to be a preservation abnormality, and a second type of abnormality information feedback is generated, which includes the water loss rate value Wr and its exceeding data; If Br > Bs or Lr > Ls, it is judged as a quality abnormality, and a third type of abnormality information feedback is generated, which includes the browning index Br or the browning area ratio Lr and its exceeding data. If Sr < Ss, it is judged as a quality abnormality, and a fourth type of abnormality information feedback is generated, which includes the morphological score Sr and its substandard data; If Er > Es, it is determined to be an environmental anomaly, and a fifth type of anomaly information feedback is generated, which includes the ethylene concentration Er and its exceedance data; S400: Based on all the judgment results, generate a structured comprehensive quality assessment report for broccoli.

[0010] Preferably, the intelligent decision execution unit receives the information feedback from the analysis feedback module and generates the following control decisions: Upon receiving feedback of the first type of abnormal information, a decision on moldy isolation and recycling is generated; Upon receiving feedback of the second type of abnormal information, a precise humidity and temperature control decision is generated; Upon receiving feedback of the third type of abnormal information, a controlled atmosphere regulation or sorting decision is generated; Upon receiving feedback of the fourth type of anomaly, an optimized storage strategy decision is generated. Upon receiving feedback of the fifth type of abnormal information, a gas composition refresh or ethylene adsorption decision is generated. The intelligent decision execution unit transmits the generated control decisions to the warehouse environment control module.

[0011] Preferably, the early warning module is used to transmit the comprehensive quality assessment report and recommended measures of broccoli to a remote terminal.

[0012] The technical effects and advantages of this invention are as follows: 1. This invention integrates an online multimodal quality detection module, an analysis and feedback module, an intelligent decision-making and execution unit, and an execution mechanism to construct a complete preservation feedback control loop. At the same time, this invention can automatically perform online full inspection of key indicators such as mold, water loss rate, color, and ethylene concentration, and automatically trigger differentiated preservation measures such as precise humidity and temperature adjustment, modified atmosphere packaging, and isolation treatment based on the judgment results. This changes the previous extensive management mode that relied on manual spot checks and passive responses. It can detect problems in real time during the storage process and actively intervene to reduce losses and extend shelf life. 2. In response to different deterioration problems of broccoli, the system of this invention will take different treatment methods: batches that have wilted due to dehydration but have not become moldy will be automatically guided to a precise humidity and temperature control zone for rescue; batches that have begun to yellow or have accumulated ethylene will be automatically treated with modified atmosphere or repackaged; and batches that are moldy or severely rotten will be immediately isolated. This can maximize the preservation of the commercial value of broccoli and reduce unnecessary losses. Attached Figure Description

[0013] Figure 1 This is an overall flowchart of the intelligent management system of the present invention. Detailed Implementation

[0014] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent broccoli storage and preservation environment control system involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Reference Figure 1 As shown, the present invention provides an intelligent warehousing and preservation environment control system for broccoli, including an intelligent management system. The intelligent management system includes a warehousing environment control module, an online multimodal quality detection module, an analysis and feedback module, an intelligent decision execution unit, and an early warning and reminder module. The storage environment control module is used to automatically control environmental parameters such as temperature, humidity, and gas composition in the broccoli storage warehouse and to set target parameters. The online multimodal quality inspection module is set up in the receiving area or between storage shelves in the warehouse to perform real-time, non-destructive inspection on broccoli in storage or circulation in order to collect its multidimensional quality data. The analysis and feedback module is connected to the online multimodal quality detection module. It is used to receive and analyze multidimensional quality data and generate information feedback on the broccoli quality and safety status assessment results. The intelligent decision execution unit is connected to the analysis feedback module and the warehouse environment control module. It is used to generate corresponding control decisions based on information feedback and transmit the control decisions to the warehouse environment control module. The warehouse environment control module drives the execution mechanism set inside it to perform environmental regulation or physical operation according to the control decisions, so as to intervene or isolate the broccoli batches with abnormal quality or safety risks. The early warning and reminder module is connected to the intelligent decision-making and execution unit and the analysis and feedback module, and is used to send information feedback and control decisions to the remote terminal.

[0016] In this application embodiment, quality abnormalities include: mold growth, excessive water loss, abnormal browning, and abnormal ethylene concentration.

[0017] In this embodiment of the application, the remote terminal specifically includes, but is not limited to, at least one of the following: (1) Monitoring computers and large displays deployed in the factory's central control room are used to centrally display the production status of the entire production line and real-time alarms; (2) The mobile communication devices of production line supervisors and quality control personnel can receive push messages through dedicated applications (APP), SMS or instant messaging tools to achieve mobile monitoring and real-time response; (3) An interface for connecting to a cloud server, used for long-term storage and big data analysis of warning logs and processing records; The aforementioned remote terminal deployment methods can be used individually or in combination to build a comprehensive information channel from the field to the remote, and from fixed to mobile terminals, ensuring that quality anomalies can be detected and handled in a timely manner.

[0018] Reference Figure 1 As shown, the present invention provides an intelligent warehousing and preservation environment control system for broccoli. The warehousing environment control module includes an environment control sub-module set in each area of ​​the broccoli storage warehouse, which is used to control the environment of each area to operate according to the set target parameters. The system also includes a main storage flow channel, wherein the actuators are connected in parallel or bypassed with the main storage flow channel, and the actuators include: The precise humidity and temperature control submodule is used to perform localized humidification and precise cooling treatment on batches of broccoli that have excessive water loss or are at risk of wilting. The modified atmosphere packaging and gas control submodule is used to automatically adjust the gas composition inside the packaging or in a local space based on ethylene concentration and respiration intensity data, so as to carry out intelligent modified atmosphere preservation. The color and mold intelligent sorting submodule is used to integrate hyperspectral and image data to identify areas of abnormal browning or mold, and to intelligently sort and isolate broccoli. The intelligent isolation and recycling submodule is used to automatically transfer, isolate, and dispose of abnormal batches of broccoli that show clear signs of mold or severe spoilage.

[0019] In this embodiment of the application, the intelligent decision-making execution unit is further configured to: plan an ordered processing path for each batch of abnormalities in the execution mechanism based on the quality anomaly type provided by the analysis feedback module, and control each sub-module of the execution mechanism to run sequentially according to this path. The planning of the processing path follows these steps: S11. If mold is detected, regardless of whether other abnormalities exist at the same time, the batch will be directly routed to the intelligent isolation and recycling submodule. S12. If no mold is detected and the water loss rate exceeds the standard, the system will be routed to the precision humidity and temperature control submodule for priority processing, and a decision will be made based on the processing results whether to enter the subsequent submodule or return to the main storage area. S13. If no mold growth or excessive water loss is detected, but there is abnormal browning or abnormal ethylene concentration, then the route will be planned to the modified atmosphere packaging and gas control submodule. S14. If only local color abnormalities are detected and there is no mold, then plan to route it to the color and mold intelligent sorting submodule for sorting, and return the qualified part to storage. All broccoli batches processed through the planning paths S12 to S14 are ultimately routed to the appropriate storage area that meets their quality requirements for subsequent preservation.

[0020] In this embodiment of the application, as a preferred implementation of the intelligent isolation and recycling submodule, it includes a warehouse handling robot, an isolation bin, a non-conforming product area, and a supporting control system. When the system determines that a batch is abnormally moldy (Class I abnormality), the intelligent decision execution unit generates a warehouse transfer instruction containing the coordinates of the target storage location and the destination. After receiving instructions, warehouse handling robots (such as AGV automated guided vehicles or composite robots with robotic arms) autonomously navigate to the target location, retrieve the problematic pallets or boxes, and transport them to the designated isolation area or non-conforming goods temporary storage area. Throughout the process, the early warning module simultaneously sends a notification that "abnormal batches have been isolated" and disposal suggestions (such as scrapping or downgrading) to the management terminal. This design enables a rapid, unmanned response to safety hazards, avoids cross-contamination, and improves the level of automation in warehouse management.

[0021] Reference Figure 1As shown, this invention provides an intelligent warehousing and preservation environment control system for broccoli. The online multimodal quality detection module integrates multiple sub-units for collecting multidimensional quality data of broccoli, wherein the multiple sub-units include: The near-infrared spectroscopy detection subunit emits near-infrared light in a specific band and receives the diffuse reflectance spectrum of broccoli. It calculates and outputs the water loss rate value Wr of broccoli in real time through the built-in spectral analysis model. The hyperspectral imaging subunit scans the surface of broccoli online to acquire high-dimensional image cube data containing continuous spectral information. Through spectral feature matching and classification algorithms, it simultaneously extracts the feature index Mr, the color browning index Br, and the relative chlorophyll content Cr from the data to characterize the risk of mold. The machine vision image acquisition subunit acquires high-definition color images of broccoli under uniform lighting conditions. It analyzes the compactness of the florets, color uniformity, and surface defects through image processing algorithms, and then calculates the morphological score Sr and the proportion of local browning area Lr. The gas sensing subunit monitors the ethylene concentration Er, carbon dioxide concentration CO2r, and relative humidity RHr in the broccoli storage microenvironment. The near-infrared spectroscopy detection subunit, hyperspectral imaging subunit, machine vision image acquisition subunit, and gas sensing subunit are distributed and deployed at key nodes in the warehouse area, including inbound inspection stations, storage shelf areas, and outbound verification points, to inspect broccoli in static storage or during turnover. Each subunit is started by the system according to the inspection task, and a unified data association mechanism ensures that the multi-dimensional inspection data of the same batch of broccoli is consistent in time and space.

[0022] The analysis feedback module is configured to perform the following analysis steps to process the multidimensional quality data collected from the online multimodal quality inspection module. The specific process is as follows: S100, Data Reception: Receive and associate multidimensional quality data from the online multimodal quality detection module. The data includes at least the water loss rate Wr, mold characteristic index Mr, browning index Br, chlorophyll content Cr, morphology score Sr, browning area percentage Lr, and ethylene concentration Er. S200, Threshold Comparison: Compare the received quantitative data with the corresponding quality thresholds preset in the database, including: comparing Wr with the water loss threshold Ws, comparing Mr with the mold threshold Ms, comparing Br with the browning threshold Bs, comparing Sr with the morphology threshold Ss, comparing Lr with the local browning threshold Ls, and comparing Er with the ethylene threshold Es. S300. Perform intelligent judgment based on the comparison results: If Mr > Ms, it is judged as a safety anomaly, and the first type of anomaly information feedback is generated, which includes the mold characteristic index Mr and its exceeding data. If Wr > Ws, it is determined to be a preservation abnormality, and a second type of abnormality information feedback is generated, which includes the water loss rate value Wr and its exceeding data; If Br > Bs or Lr > Ls, it is judged as a quality abnormality, and a third type of abnormality information feedback is generated, which includes the browning index Br or the browning area ratio Lr and its exceeding data. If Sr < Ss, it is judged as a quality abnormality, and a fourth type of abnormality information feedback is generated, which includes the morphological score Sr and its substandard data; If Er > Es, it is determined to be an environmental anomaly, and a fifth type of anomaly information feedback is generated, which includes the ethylene concentration Er and its exceedance data; S400: Based on all the judgment results, generate a structured comprehensive quality assessment report for broccoli as feedback.

[0023] In this embodiment, the preset quality thresholds at each level in the analysis feedback module are comprehensively set based on the post-harvest physiological characteristics of broccoli, food safety regulations, industry preservation standards, and target market commodity grade requirements. These thresholds constitute the basis for the system to make intelligent judgments. The specific settings are as follows: Water loss rate threshold (Ws): set according to post-harvest physiological research on broccoli and commodity appearance requirements. Water loss rate is a key indicator that leads to wilting and weight loss; (for example, Ws can be set to 5% (percentage of fresh weight loss). When the detected water loss rate Wr > 5%, it is determined that the water loss exceeds the standard and moisturizing treatment is required).

[0024] Mold Determination Threshold (Ms): This is set based on the output value of a deep learning model that uses a hyperspectral imaging subunit to analyze the spectral characteristics of typical moldy samples (such as gray mold). The model normalizes the moldy characteristics into a risk index Mr, which is usually in the range of 0-1. In conjunction with food safety requirements, a conservative safety threshold is set. (For example, if Ms is 0.03, when Mr > 0.03, the system considers there to be an unacceptable risk of mold and determines it as a safety anomaly).

[0025] Browning thresholds (Bs, Ls): Browning is a direct manifestation of the yellowing and aging of broccoli. Bs is the overall browning index threshold, set based on changes in chromaticity coordinates; Ls is the threshold for the proportion of local lesions or browned areas; (for example, Bs can be set to a specific color difference value, or Ls can be set to 5% of the surface area of ​​a single floret. When Br > Bs or Lr > 5%, it is judged as an abnormal color).

[0026] Morphology score threshold (Ss): Morphology score Sr comprehensively reflects the appearance and commodity quality of the flower head, such as firmness and integrity. The threshold is set according to the requirements of the market for premium products: (For example, Ss is set to 80 points (out of 100). When Sr < 80, it is judged as morphology not meeting the standard.)

[0027] Ethylene concentration threshold (Es): Ethylene is a key plant hormone that accelerates the senescence of broccoli. The threshold is set based on the broccoli's sensitivity to ethylene and the optimal preservation gas conditions; (for example, Es can be set to 0.1 ppm, and when Er > 0.1 ppm, it is considered an abnormal ethylene concentration, requiring the initiation of ethylene removal or gas refresh measures). It should be noted that the specific values ​​of the above thresholds (Ws, Ms, Bs, Ls, Ss, Es) can be flexibly configured and adjusted in the system's management interface according to the broccoli variety, harvest maturity, target storage period, and customer customization needs. In this application, the above threshold ranges are provided by default as recommended values ​​based on industry research and standards, but authorized users are allowed to modify them according to actual conditions, thereby enhancing the overall adaptability and accuracy of the intelligent control system. In the above context, "deep learning model" refers to an algorithm engine deployed within a hyperspectral imaging subunit or an edge computing device connected to it. Specifically, it is a pre-trained and re-trained deep convolutional neural network or a variant thereof, specifically designed for processing hyperspectral image cubic data; its key technical features are as follows: Model architecture: It adopts an advanced architecture that can process spatial and spectral information simultaneously, such as 3D convolutional neural networks and spectral-spatial joint attention networks, to accurately extract the subtle texture features of mold spots and their unique spectral characteristics, and distinguish them from normal tissue.

[0028] Training data and process: Supervised learning was conducted using a large number of labeled hyperspectral images of broccoli covering different varieties, different mold types (such as gray mold and sclerotinia rot), and different disease severity, so that the model learned to map high-dimensional image data to a mold risk index Mr (usually normalized to the 0-1 range).

[0029] Model Output: For real-time collected broccoli surface data, the model outputs the mold risk value for each pixel or analysis area. Based on this, the system statistically analyzes and generates a mold characteristic index Mr representing the risk of the entire batch or a single floret.

[0030] Through this model, the system achieves highly sensitive and automated detection of early mold growth, especially internal or early-stage mold growth that is difficult to detect with the naked eye; In this embodiment of the application, the calculation process for the morphological score Sr is as follows: D100, Image Acquisition: The machine vision image acquisition subunit acquires images of broccoli florets under uniform lighting conditions using high-definition color cameras deployed between shelves or on inspection devices. D200, Image Preprocessing and Target Segmentation: The original image is standardized (e.g., color space conversion, noise reduction), and based on the color and texture features of broccoli, the flower head region is accurately extracted from the background using image segmentation algorithms (e.g., color thresholding combined with edge detection) as the object of subsequent analysis; D300, Feature Parameter Extraction and Normalization: Analyze the segmented flower ball region and extract original features of multiple dimensions for quantitative evaluation of its commerciality. In order to eliminate the differences in the units and numerical ranges of each original feature, it is necessary to normalize it, that is, to convert it to a uniform comparable scoring range (e.g., 0 to 100 points) through a preset mapping rule. The mapping rules can be configured based on the data distribution of the features and their positive or negative correlation with quality. The parameters obtained after processing are as follows: Compactness parameter Q1: This parameter is a quantitative indicator of the compactness and regularity of the flower head structure. It is obtained by calculating the roundness of the flower head area or its area filling rate (i.e., the ratio of the actual area of ​​the flower head to the area of ​​its smallest bounding rectangle). The original feature value is then normalized as described above to obtain parameter Q1. The higher the value, the better the compactness. Integrity parameter Q2: This parameter is a quantitative indicator of physical damage and structural defects on the surface of the flower head. It is obtained by analyzing the convexity defects of its contour (identifying and measuring abnormal depressions on the contour line) or calculating the smoothness of the contour (assessing abrupt changes in local curvature) to obtain the original features characterizing the degree of defects. The original feature value is then normalized as described above to obtain parameter Q2. The lower the value, the worse the integrity. Specification uniformity parameter Q3: This parameter is a quantitative indicator of the consistency of flower head size within the same batch. It is obtained by statistically analyzing the dispersion of the specifications of all individual flower heads within the batch: First, obtain the specification data of each flower head in the batch (such as pixel area or equivalent diameter); then, calculate the statistical distribution characteristics of these data (such as coefficient of variation, standard deviation, etc.) as the original dispersion index. After normalizing the original dispersion index as described above, the parameter Q3 is obtained. The smaller the value, the higher the uniformity. D400. The normalized set of parameters (Q1, Q2, Q3) obtained in step D300 is comprehensively calculated according to the preset fusion rules to generate the final morphological score Sr. The fusion rules aim to reflect the contribution weights Ω1, Ω2, and Ω3 of each dimension feature to the overall morphological quality (where Ω1, Ω2, and Ω3 are the weight ratios of Q1, Q2, and Q3 respectively, and Ω1+Ω2+Ω3=1). The form includes, but is not limited to, weighted summation, rule-based decision tree, or machine learning model mapping. The morphological score Sr intuitively represents the overall morphological quality of broccoli in terms of firmness, integrity, and uniformity of size. The higher the score, the better the overall morphological quality. The formula for calculating the morphological score Sr is as follows: .

[0031] Reference Figure 1 As shown, this invention provides an intelligent warehousing and preservation environment control system for broccoli. The intelligent decision execution unit receives information feedback from the analysis feedback module and generates the following control decisions: Upon receiving feedback of the first type of abnormal information, a decision on moldy isolation and recycling is generated; Upon receiving feedback of the second type of abnormal information, a precise humidity and temperature control decision is generated; Upon receiving feedback of the third type of abnormal information, a controlled atmosphere regulation or sorting decision is generated; Upon receiving feedback of the fourth type of anomaly, an optimized storage strategy decision is generated. Upon receiving feedback of the fifth type of abnormal information, a gas composition refresh or ethylene adsorption decision is generated. The intelligent decision execution unit transmits the generated control decisions to the warehouse environment control module.

[0032] The early warning and reminder module is used to transmit the comprehensive quality assessment report and recommended measures for broccoli to the remote terminal.

[0033] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart storage and preservation environment control system for broccoli, characterized in that: The system comprises an intelligent management system, which comprises a warehouse environment control module, an online multi-modal quality detection module, an analysis feedback module, an intelligent decision execution unit and an early warning reminding module. The warehouse environment control module is used for automatically controlling and setting target parameters of temperature, humidity and gas composition environment parameters in the broccoli warehouse; The online multi-modal quality detection module is arranged in the storage area or between the storage shelves of the warehouse, and is used for real-time and non-destructive detection of the stored or transferred broccoli to collect multi-dimensional quality data thereof; The analysis feedback module is communicatively connected to the online multi-modal quality detection module, and is used for receiving and analyzing the multi-dimensional quality data, and generating information feedback of the quality and safety state evaluation results of the broccoli; The intelligent decision execution unit is communicatively connected to the analysis feedback module and the warehouse environment control module, and is used for generating corresponding control decisions according to the information feedback, and transmitting the control decisions to the warehouse environment control module, so that the warehouse environment control module drives the execution mechanism arranged therein to perform environmental regulation or physical operation according to the control decisions, so as to intervene or isolate and dispose the batches of broccoli with quality abnormalities or safety risks; The early warning reminding module is communicatively connected to the intelligent decision execution unit and the analysis feedback module, and is used for sending the information feedback and the control decisions to a remote terminal.

2. The Brassica oleracea intelligent storage and fresh-keeping environment control system according to claim 1, characterized in that: The quality abnormalities include mold, excessive water loss rate, abnormal color browning and abnormal ethylene concentration.

3. The Brassica oleracea intelligent storage and fresh-keeping environment control system according to claim 2, characterized in that: The warehouse environment control module comprises environment control sub-modules arranged in each area of the broccoli warehouse, which are used for controlling the environment of each area to operate according to the set target parameters; The system further comprises a main warehouse flow channel, wherein the execution mechanism is arranged in parallel or bypass with the main warehouse flow channel, and the execution mechanism comprises: A precise humidity and temperature adjusting sub-module is used for locally humidifying and precisely cooling the batches of broccoli with excessive water loss rate or at risk of wilting; An atmosphere packaging and gas regulating sub-module is used for automatically adjusting the gas composition in the packaging or local space according to the ethylene concentration and respiration intensity data, and performing intelligent atmosphere preservation; A color and mold intelligent sorting sub-module is used for fusing hyperspectral and image data to identify the color abnormal browning or mold area, and intelligently sorting and isolating the broccoli; An intelligent isolation and recycling sub-module is used for automatically moving, isolating and disposing the abnormal batches of broccoli with obvious mold or serious corruption.

4. The Brassica oleracea intelligent storage and fresh-keeping environment control system according to claim 3, characterized in that: The intelligent decision execution unit is further configured to plan an ordered processing path of each abnormal batch in the execution mechanism according to the quality abnormality type provided by the analysis feedback module, and control the sub-modules of the execution mechanism to operate in sequence according to the path: The planning of the processing path follows the following steps: S11, if mold is detected, regardless of whether other abnormalities exist simultaneously, the batch is directly routed to the intelligent isolation and recycling sub-module; S12, if no mold is detected and the water loss rate is excessive, the batch is routed to the precise humidity and temperature adjusting sub-module for priority processing, and whether to enter the subsequent sub-module or return to the main storage area is determined according to the processing result; S13, if no mold and excessive water loss rate are detected, but color abnormal browning or ethylene concentration abnormality exists, then planning to route it to the modified atmosphere packaging and gas control submodule; S14, if only local color abnormality is detected and no mold exists, then planning to route it to the color and mold intelligent sorting submodule for sorting, and the qualified part is returned to storage; All batches of broccoli processed by the above S12 to S14 planning path are finally routed to the corresponding storage area according to their quality state for subsequent preservation.

5. The Brassica oleracea intelligent warehousing and fresh-keeping environment control system according to claim 1, characterized in that: The online multi-modal quality detection module integrates multiple groups of subunits for collecting multi-dimensional quality data of broccoli, wherein the multiple groups of subunits include: A near-infrared spectrum detection subunit emits near-infrared light of a specific wave band and receives the diffuse reflection spectrum of the broccoli, and calculates and outputs the water loss rate Wr of the broccoli in real time through a built-in spectrum analysis model; A hyperspectral imaging subunit online scans the surface of the broccoli to obtain high-dimensional image cube data containing continuous spectral information, and extracts feature indexes Mr, color browning index Br and relative chlorophyll content Cr for representing mold risk from the data through spectral feature matching and classification algorithms; A machine vision image acquisition subunit acquires high-definition color images of the broccoli under uniform illumination conditions, analyzes the tightness, color uniformity and surface defects of the flower ball through image processing algorithms, and then calculates the morphology score Sr and local browning area ratio Lr; A gas sensing subunit monitors the ethylene concentration Er, carbon dioxide concentration CO2r and relative humidity RHr in the storage microenvironment of the broccoli.

6. The Brassica oleracea intelligent warehousing and fresh-keeping environment control system according to claim 5, characterized in that: The near-infrared spectrum detection subunit, hyperspectral imaging subunit, machine vision image acquisition subunit and gas sensing subunit are distributedly arranged at key nodes of the warehouse area, including the warehouse entry detection station, storage shelf interval and warehouse exit review point, for detecting the broccoli in static storage or in transit; each subunit is started by the system according to the detection task, and the unified data correlation mechanism is used to ensure that the multi-dimensional detection data of the same batch of broccoli is consistent in time and space.

7. The Brassica oleracea intelligent warehousing and fresh-keeping environment control system according to claim 5, characterized in that: The analysis feedback module is configured to perform the following analysis steps to process the multi-dimensional quality data collected by the online multi-modal quality detection module, and the specific process is as follows: S100, data receiving: receiving and correlating the multi-dimensional quality data from the online multi-modal quality detection module, the data at least including water loss rate Wr, mold feature index Mr, color browning index Br, chlorophyll content Cr, morphology score Sr, browning area ratio Lr, and ethylene concentration Er; S200, threshold comparison: comparing each quantitative data received with the corresponding quality threshold preset in the database, including: comparing Wr with the water loss rate threshold Ws, comparing Mr with the mold threshold Ms, comparing Br with the browning threshold Bs, comparing Sr with the morphology threshold Ss, comparing Lr with the local browning threshold Ls, and comparing Er with the ethylene threshold Es; S300, performing intelligent decision according to the comparison result: If Mr>Ms, it is determined that there is a safety abnormality, and a first type of abnormality information feedback is generated, which contains the mold characteristic index Mr and its exceeding data; If Wr>Ws, it is determined that there is a preservation abnormality, and a second type of abnormality information feedback is generated, which contains the water loss rate Wr and its exceeding data; If Br>Bs or Lr>Ls, it is determined that there is a quality abnormality, and a third type of abnormality information feedback is generated, which contains the color browning index Br or the browning area proportion Lr and its exceeding data; If Sr<Ss, it is determined that there is a quality abnormality, and a fourth type of abnormality information feedback is generated, which contains the morphology score Sr and its substandard data; If Er>Es, it is determined that there is an environmental abnormality, and a fifth type of abnormality information feedback is generated, which contains the ethylene concentration Er and its exceeding data; S400, comprehensively judge the quality of broccoli, and generate a structured comprehensive evaluation report of broccoli quality.

8. The Brassica oleracea intelligent warehousing and fresh-keeping environment control system according to claim 7, characterized in that: The intelligent decision execution unit receives the information feedback sent by the analysis feedback module, and generates the following control decisions: Receive the first type of abnormality information feedback and generate mold isolation and recycling decisions; Receive the second type of abnormality information feedback and generate precise humidity and temperature adjustment decisions; Receive the third type of abnormality information feedback and generate modified atmosphere control or sorting decisions; Receive the fourth type of abnormality information feedback and generate optimized storage strategy decisions; Receive the fifth type of abnormality information feedback and generate gas component refreshing or ethylene adsorption decisions; The intelligent decision execution unit transmits the generated control decisions to the warehouse environment control module.

9. The Brassica oleracea intelligent warehousing and fresh-keeping environment control system according to claim 7, characterized in that: The early warning reminding module is used to send the comprehensive evaluation report of broccoli quality and the recommended measures to the remote terminal.

Images