Shipborne squid quality traceability intelligent terminal device, system and method

Abstract

The invention discloses shipborne squid quality traceability intelligent terminal equipment, system and method. The method comprises the following steps: a data acquisition module acquires image and label information of a squid woven bag through an RFID and a visual sensor; the environment sensing module dynamically adjusts the brim and multi-light-source illumination, and optimizes the image quality; the visual identification processing module uses a deep learning algorithm to accurately identify size grades, and temperature data are fused to form preliminary fusion data; the central control processing module integrates RFID, size grade, temperature and satellite positioning data to realize time sequence synchronization and data fusion; and the communication module uploads data to the cloud block chain platform in real time through a multi-redundant network and provides multi-modal feedback. According to the scheme, the RFID tag identification accuracy, the visual identification stability, the data transmission reliability and the equipment environment adaptability are remarkably improved, and the efficiency and the reliability of full-link traceability of squid fishing are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the intersection of Internet of Things technology and aquatic product supply chain management, and in particular to a shipborne squid quality traceability intelligent terminal device, system and method. Background Art

[0002] Currently, radio frequency identification (RFID) and visual recognition technologies are widely used in deep-sea squid fishing to achieve quality traceability and refined management of squid. However, traditional traceability equipment and methods still have many problems and shortcomings, as shown below:

[0003] 1. RFID tag recognition is prone to collisions and misreading: Existing traceability systems mostly use a single-band, single-antenna architecture. Due to the lack of effective collaborative channels between rotating passive RFID tags, tag signal collisions are prone to occur during batch recognition. When the spatial layout between the RF antenna and the tag is unreasonable or the antenna power is not precisely controlled, misreading and missed readings are frequent, seriously affecting data accuracy and traceability effectiveness.

[0004] 2. Visual recognition has poor environmental adaptability and a high misjudgment rate: The text on the surface of squid woven bags is easily obscured by dirt, water stains, reflections, or shadows in open-sea environments. Existing camera technology lacks effective adaptive fill lighting and ambient light adjustment, resulting in image quality fluctuations and significantly reduced visual recognition accuracy. Furthermore, traditional visual recognition algorithms struggle to efficiently and accurately identify text in complex environments, severely hindering the automated implementation of size classification.

[0005] 3. Isolated multi-source data collection, poor data fusion and synchronization: Traditional equipment often collects data such as size, temperature, and location separately. Each data source is stored and transmitted in isolation, lacking a unified timing alignment and data fusion mechanism. This results in data silos and information fragmentation, making it difficult to achieve complete end-to-end data traceability and efficient data tracking.

[0006] 4. Insufficient equipment environmental adaptability and poor reliability: The offshore fishing environment is harsh, characterized by high humidity, high salt spray, large temperature differences, and severe hull vibration. Existing traceability equipment lacks targeted protection and shock absorption measures, resulting in poor long-term operational stability and short lifespan of the equipment, making it difficult to ensure the reliability of data collection.

[0007] 5. Poor data transmission stability in open ocean areas: Communication signal coverage in high seas is unstable. Traditional single-use wireless communication methods (such as relying solely on Wi-Fi or a single cellular network) are prone to data upload interruptions, making it difficult to ensure data integrity and real-time performance. Furthermore, effective anti-tampering measures are not implemented during data transmission and storage, posing data security risks.

[0008] 6. The manual intervention mode has a high error rate and lacks an effective feedback mechanism: The operating buttons in the manual mode of existing equipment are simply designed, which is prone to false triggering or repeated input. In addition, the single visual feedback or sound feedback method is difficult to effectively prompt operators in the high-noise and complex shipboard environment, affecting data entry accuracy and work efficiency.

[0009] In view of this, the present invention proposes a shipborne squid quality traceability intelligent terminal device, system and method. Summary of the Invention

[0010] The purpose of the present invention is to provide a shipborne squid quality traceability intelligent terminal device, system and method to accurately detect defects and effectively improve the packaging quality and reliability of IGBT modules.

[0011] In a first aspect, the present invention provides an intelligent method for tracing the quality of squid on board a ship, comprising:

[0012] S101: Capturing a first target image within a target area, reading RFID tag information of the first target, and marking the first target image and the corresponding RFID tag information as initial data;

[0013] S102: The automatic brim adjustment mechanism based on dynamic ambient light sensing dynamically adjusts the extension and retraction state of the brim, and adjusts the illumination angle and brightness in real time through the multi-light source lighting system to optimize the size feature image data of the first target image;

[0014] S103: Inputting the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and generating preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor;

[0015] S104: Synchronously input the RFID tag information, preliminary fusion data, and fishing location data, and construct a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules;

[0016] S105: Upload the full-link traceability data packet to the cloud data center in real time, and trigger the multimodal interactive feedback device at the same time to provide real-time feedback on the data binding status in the form of sound, light and touch.

[0017] As a preferred technical solution of the first aspect of the present invention, the automatic adjustment mechanism of the brim based on dynamic ambient light sensing includes:

[0018] The ambient illumination value is collected in real time through the light sensor probe installed on the top of the device;

[0019] When the ambient illumination is lower than the ambient illumination threshold, the electric actuator is controlled to drive the brim to automatically retract;

[0020] When the ambient illumination is higher than the ambient illumination threshold, the electric actuator is controlled to drive the brim of the hat to automatically extend;

[0021] Based on the ambient light values ​​collected in real time by the light sensor, the light source intensity is dynamically adjusted and the light source angle is controlled to maintain constant image brightness within the camera's field of view.

[0022] As a preferred technical solution of the first aspect of the present invention, the deep convolutional network model includes:

[0023] Extract local and overall features of historical target size images to generate multi-scale image feature data; construct a deep convolutional network containing multi-scale convolution kernels and train it based on historical target size graded image data. Repeat the training until the model's accuracy in target size recognition reaches the preset standard, and determine the convolution kernel parameters;

[0024] The optimized target size feature image is input into the deep convolutional network model, and the corresponding size classification label is output according to the preset size classification standard.

[0025] As a preferred technical solution of the first aspect of the present invention, the data fusion algorithm and timestamp synchronization rules include:

[0026] Calculate the time difference between the collection time of RFID, visual, temperature and positioning data and the fishing reference time;

[0027] Data with a time difference less than the preset time threshold is considered valid synchronization data and is timestamped in a unified format.

[0028] Using the timestamp of fishing as a benchmark, the data from each sensor are time-aligned to form a time-synchronized squid traceability data packet.

[0029] As a preferred technical solution of the first aspect of the present invention, the multimodal interactive feedback device real-time feedback of the data binding status includes:

[0030] When the central control processing module confirms that the data packet is complete and valid, the control status indicator light turns green and a short sound prompt is triggered at the same time;

[0031] When the central control processing module detects missing or incorrect data, the control status indicator light turns red, accompanied by a warning sound and vibration feedback, to remind staff to review the data binding status.

[0032] In a second aspect, the present invention provides a shipborne intelligent squid quality traceability system. Based on the implementation of the first aspect, the system is integrated into an embedded circuit board within a terminal device. The PCB includes a data acquisition module, an environmental sensing module, a visual recognition processing module, a central control processing module, and a communication module. Each module is connected via wired and / or wireless communication to achieve real-time data binding and cloud synchronization for the entire squid catching, sorting, and packaging process.

[0033] The data acquisition module collects a first target image in the target area, reads the RFID tag information of the first target, and marks the first target image and the corresponding RFID tag information as initial data; the data acquisition module sends the initial data to the central control processing module;

[0034] The environmental sensing module uses an automatic brim adjustment mechanism based on dynamic ambient light sensing to dynamically adjust the brim's extension and contraction state. It also uses a multi-light source lighting system to adjust the lighting angle and brightness in real time, optimizing the dimensional feature image data of the first target image. This dimensional feature image data is then fed into the visual recognition processing module.

[0035] The visual recognition processing module inputs the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and simultaneously generates preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor; and sends the preliminary fusion data to the central control processing module;

[0036] The central control processing module synchronously inputs RFID tag information, preliminary fusion data, and fishing location data, and constructs a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules;

[0037] The communication module uploads the full-link traceability data packet to the cloud data center in real time, and at the same time triggers the multimodal interactive feedback device to provide real-time feedback on the data binding status in the form of sound, light and touch.

[0038] As a preferred technical solution of the second aspect of the present invention, the data acquisition module includes an automatic acquisition mode and a manual acquisition mode;

[0039] In automatic acquisition mode, a high-resolution industrial camera equipped with a polarizing filter and an adaptive fill light module can identify the size markings on the squid woven bags in real time, read the label codes, and automatically associate them with GPS or Beidou positioning data to generate a unified data package.

[0040] In manual collection mode, the squid size information is manually input through a physical button with anti-accidental touch logic, triggering the reading and binding of RFID tags. The bound data generates a data packet with a marked field, and the manual operation status is prompted through a multimodal feedback device.

[0041] As a preferred technical solution of the second aspect of the present invention, the method for generating the preliminary fusion data is:

[0042] The target size classification information is compared with the collection timestamp corresponding to the quick-freezing chamber temperature information, and structured data fusion is performed when the time difference between the two is less than the preset threshold; when the credibility of the size classification information or the temperature information abnormality exceeds the set threshold, the multimodal feedback device alarm is automatically triggered and the manual review mode is started.

[0043] In a third aspect, the present invention provides an intelligent method for tracing the quality of squid on board a ship, which is used to integrate and fix the second aspect, including a conveyor rolling track fixed on an adjustable height bracket, wherein an anti-corrosion shell is provided on the side of the adjustable height bracket where the conveyor rolling track is located, and the anti-corrosion shell adopts an IP68 protection grade industrial anti-corrosion material, and the anti-corrosion shell includes:

[0044] RFID identification antenna, used to cooperate with the RFID reader, transmit radio frequency signals to activate the passive RFID tag in the squid woven bag, and receive the unique coding information returned by the tag;

[0045] Display screen, showing equipment operating status, data binding status, real-time temperature, size grade display, and providing a manual interaction interface in manual mode;

[0046] Button, used for the operator to quickly select the squid size class by pressing a button in manual mode;

[0047] Status indicator light, which displays the device operating status and data binding results in real time:

[0048] The light sensor monitors the ambient light intensity in real time and sends the light intensity to the light sensing module to control the automatic extension and retraction of the device's brim and the automatic adjustment of the brightness of multiple light sources.

[0049] Positioning antenna, realize GPS / Beidou satellite positioning, obtain the squid fishing location, and transmit the squid fishing location data.

[0050] As a preferred technical solution for the third aspect of the present invention, the RFID identification antenna adopts a "three-section" identification antenna, and RFID radio frequency antenna segments are respectively installed in the first two sections of the three-section structure; the two RFID radio frequency antenna segments form an angle of approximately 30 degrees in the spatial layout, presenting a front-to-back staggered and angled setting, forming a spatially differentiated identification channel.

[0051] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0052] The solution of the present invention significantly improves the recognition accuracy, data integrity, communication reliability, equipment stability and human-computer interaction efficiency of ship-borne squid quality traceability equipment, meets the actual needs of full-link data traceability in the complex environment of distant-water fishing, and provides strong technical support for the intelligent and refined management of distant-water fishing. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0054] Figure 1 This is a structural framework diagram of the intelligent terminal device for tracing the quality of ocean squid according to the present invention;

[0055] Figure 2 This is a schematic diagram of the framework flow of the intelligent system for tracing the quality of pelagic squid of the present invention.

[0056] In the figure: 1. RFID identification antenna; 2. Display screen; 3. Button; 4. Status indicator light; 5. Light sensor; 6. Positioning antenna; 7. Conveyor rolling track; 8. Adjustable height bracket. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the implementation of this application clearer, the technical solutions in the implementation of this application will be described in more detail below in conjunction with the drawings in the implementation of this application.

[0058] In the accompanying drawings, the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The described embodiments are part of the embodiments of the present application, rather than all of the embodiments. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be understood as limiting the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application. The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0059] Example 1

[0060] See also Figure 1 As shown, this embodiment provides an intelligent terminal device for tracing the quality of pelagic squid, including a conveyor rolling track 7 fixed on an adjustable height bracket 8. The conveyor rolling track 7 is provided with an anti-corrosion shell on the side of the adjustable height bracket 8. The anti-corrosion shell adopts an industrial anti-corrosion material with an IP68 protection grade, and the anti-corrosion shell includes:

[0061] The RFID identification antenna 1 is used to cooperate with the RFID reader to transmit radio frequency signals to activate the passive RFID tag in the squid woven bag and receive the unique coding information returned by the tag.

[0062] For example, ultra-high frequency (UHF) or high frequency (HF) antennas are used, with typical operating frequencies of 860-960 MHz or 13.56 MHz; multiple antenna arrays are configured to improve signal stability and anti-interference capabilities; and space division multiplexing and adaptive power control technologies are used to achieve fast, stable and accurate tag identification, thereby ensuring data link reliability.

[0063] Specifically, the RFID identification antenna 1 adopts a "three-segment" identification antenna, with RFID radio frequency antenna segments installed in the first two sections of the three-segment structure. The two RFID radio frequency antenna segments form an angle of approximately 30 degrees in the spatial layout, and the spatial layout presents a front-to-back staggered and angled setting, forming a spatially differentiated identification channel, effectively reducing tag collisions caused by simultaneous tag signal responses.

[0064] The installation position of the RFID radio frequency antenna segment and the movement trajectory of the RFID tag in the squid woven bag are optimized to ensure that when the woven bag loaded with squid moves from left to right on the sliding roller track, there is always a point with the shortest straight-line distance between the RFID radio frequency antenna segment and the tag, so that the recognition power at the point with the shortest straight-line distance between the antenna and the tag reaches the optimal value, avoiding tag interference caused by excessive power and misreading caused by insufficient power, and significantly improving the tag reading success rate and accuracy.

[0065] Display screen 2 displays the device operating status, data binding status, real-time temperature, and size grade, and provides a manual interaction interface in manual mode.

[0066] For example, the 10-inch industrial touch screen adopts a fully laminated capacitive panel with AG anti-glare coating. The screen brightness supports automatic adjustment of 500-1500nit (based on the ambient light sensor). In raindrop sensing mode, the hydrophobic nano-coating is triggered to be energized (contact angle > 150°), ensuring that the operation interface can still be clearly displayed in heavy rain (rainfall ≥ 50mm / h) or strong sunlight (illuminance ≥ 100,000 lux) environments; the RFID reading window is equipped with oleophobic and hydrophobic optical glass (transmittance ≥ 92%) to eliminate the impact of water stains and reflections on label reading, realize intuitive and reliable human-computer interaction, and improve user experience and data entry efficiency.

[0067] Button 3 is used in manual mode for the operator to quickly select the squid size level (such as extra large, large, medium, small, and small strips) by pressing the button. The number of buttons corresponding to the squid size level is set, and a design is provided to prevent accidental touches (such as a long physical button travel and a logic of more than 3 seconds between two inputs);

[0068] Built-in tactile feedback (micro-vibration) locks manual input in automatic mode. The manual intervention interface is activated only after the camera fails recognition three times in a row or the user presses and holds the "Override" button for five seconds. If automatic recognition fails, manual intervention ensures data accuracy. After the entered data is bound to the RFID tag, a red field is generated for secondary verification by the cloud-based quality inspection system to ensure data integrity throughout the entire process.

[0069] Status indicator 4, real-time display of device operating status and data binding results: The LED three-color light group includes green, yellow and red lights, which are controlled by the GPIO interface of the main control chip, quickly responding to status changes, providing clear and intuitive device operation feedback, and enhancing operational reliability.

[0070] Green light: data binding successful;

[0071] Yellow light: The device is on standby or processing;

[0072] Red light: data binding failed, system abnormality;

[0073] The light sensor 5 monitors the light intensity of the on-site environment in real time and sends the light intensity to the light sensing detection module to control the automatic extension and retraction of the device's brim and the automatic adjustment of the brightness of multiple light sources.

[0074] For example, an industrial-grade light intensity sensor (such as a silicon photodiode or an I 2 C-mount light sensor chip) outputs digital light intensity signals, accurately controls the brim and lighting system, and dynamically adjusts the image acquisition environment to ensure visual recognition stability and high accuracy.

[0075] Positioning antenna 6, realizes GPS / Beidou satellite positioning, obtains the squid fishing location, and transmits the squid fishing location data,

[0076] Specifically, dual-band or multi-band combination antennas ensure real-time data transmission and positioning accuracy in ocean environments, ensuring uninterrupted data links.

[0077] The conveyor rolls on track 7 to transport the woven bags loaded with squid to the target area for RFID and visual recognition; ensuring the physical position between the conveyor belt and the equipment is stable and keeping the recognition area constant.

[0078] Exemplarily, the conveyor rolling track 7 adopts a 304 stainless steel roller track, supports manual push-pull positioning (stroke 0-1.2m, positioning accuracy ±2mm) or electric servo control (receiving assembly line PLC instructions through CAN bus), and the roller surface is sprayed with a polyurethane anti-slip layer (friction coefficient μ≥0.6), and cooperates with the Hall encoder to provide real-time feedback on the conveying speed (adjustable 0-5m / s), which can adapt to the smooth transmission of woven bags under the bumpy conditions of the fishing boat deck.

[0079] Adjustable height bracket 8 adapts to the height changes of different fishing boat operating platforms, and accurately connects the equipment with the production line;

[0080] For example, the adjustable height bracket 8 has a built-in electric or pneumatic lifting mechanism; the stroke is adjustable in a range of 0-90 cm or larger, and is made of anodized aluminum alloy, with high corrosion resistance and salt spray resistance; it has a quick positioning and locking structure, which improves the adaptability and installation efficiency of the equipment and meets the requirements of various ship types.

[0081] It should also be noted that the corrosion-resistant casing also includes: a PWM speed-regulating fan and a semiconductor refrigeration plate active anti-fog heat dissipation system. When the temperature and humidity sensor detects that the relative humidity in the cabin is >85% or the temperature drops suddenly (ΔT≥10℃ / min), a forced convection cycle is started to balance the air pressure inside and outside the equipment and eliminate lens condensation; the fan air inlet is equipped with a HEPA-13 grade dust filter, which can intercept salt spray particles (particle size ≥0.3μm), ensuring that the optical components have no risk of fogging failure under working conditions of -20℃ to 50℃, and realizing all-weather high-definition operation display.

[0082] Example 2

[0083] See also Figure 2 As shown, this embodiment provides a shipborne squid quality traceability intelligent system, which is integrated into an embedded circuit board inside the terminal device. It works together through the standard interface and connection method on the PCB circuit board to jointly achieve efficient and stable operation of the quality traceability of deep-sea squid. The PCB circuit board includes a data acquisition module, an environmental sensing module, a visual recognition processing module, a central control processing module and a communication module. The modules are connected through wired and / or wireless communications to achieve real-time binding and cloud synchronization of data from the entire process of squid fishing, sorting and packaging.

[0084] The data acquisition module collects a first target image in the target area, reads the RFID tag information of the first target, and marks the first target image and the corresponding RFID tag information as initial data; the data acquisition module sends the initial data to the central control processing module;

[0085] It should be noted that the target area is a designated area on the conveyor belt. A first visual sensor and an RFID sensor are installed in the designated area. The first visual sensor and the RFID sensor are used to capture a first target image and read tag information of the first target. In this embodiment, the first target is a squid woven bag, and the squid is packed in the squid woven bag.

[0086] Specifically, the data acquisition module includes a radio frequency identification unit, an industrial computer main control unit, and an image acquisition unit.

[0087] The radio frequency identification unit uses RFID sensors to read the tag information of squid woven bags passing through a designated area. RFID tag reading, signal modulation, demodulation, encoding and decoding are all handled by this module.

[0088] The main control unit of the industrial computer is bound in real time to the company registration information, fishing vessel number and the longitude and latitude coordinate data obtained by the GPS module;

[0089] The image acquisition unit captures the squid woven bag image in real time based on the first visual sensor.

[0090] The data acquisition module includes automatic acquisition mode and manual acquisition mode, where:

[0091] Automatic acquisition mode: The device is equipped with a high-resolution CMOS industrial camera, an F1.4 large aperture fixed-focus lens, a polarizing filter, and an adaptive fill light module. It uses the YOLOv8s deep learning model to identify the squid size information printed on the woven bag in real time (such as Grade A ≥30cm, Grade B 20-30cm, Grade C <20cm), with an accuracy rate of more than 99%; the recognition result automatically triggers the UHF band RFID reader to obtain the unique code of the woven bag label, and uses the built-in industrial computer to link the company registration information, fishing vessel number and the precise coordinate data obtained by the GPS / Beidou positioning module in real time to generate a unified structured traceability data packet; the automatic acquisition mode uses a polarization filter camera and a deep learning algorithm to realize automatic identification and grading of squid size in woven bags, and dynamically binds it to the RFID tag and GPS coordinates.

[0092] Manual collection mode: A physical button with built-in anti-accidental touch logic and a three-color indicator light (green light for success, red light for failure, and yellow light for standby) allows operators to manually enter squid size information, triggering HF band RFID reading and data binding. The anti-accidental touch mechanism ensures that each collection interval is ≥3 seconds; after data entry, it is bound to the RFID tag in real time, and a red identification field is generated for secondary verification by the cloud-based quality inspection system. The manual collection mode integrates an anti-accidental touch button and a three-color indicator light, supports data verification and retransmission under manual intervention, and ensures data authenticity and integrity.

[0093] The environmental sensing module uses an automatic brim adjustment mechanism based on dynamic ambient light sensing to dynamically adjust the brim extension and contraction state, and uses a multi-light source lighting system to adjust the lighting angle and brightness in real time to optimize the size feature image data of the first target image.

[0094] It should be noted that the optimized squid size characteristic image is achieved by automatically reading the size grade markings on the woven bags, thereby efficiently achieving rapid classification and traceability management of squid sizes, rather than directly identifying the physical size of individual squids in the bags. For example: during fishing operations, the squid are placed in woven bags with clear size markings printed on the surface of the bags (for example, "large strips"); when the squid bags on the assembly line pass through the visual sensor area, the ambient light sensing module adjusts the brim position and fill light intensity in real time; after the camera captures a clear image, the CNN+OCR model accurately recognizes the size identification text on the woven bags; the size grade results are automatically output to form size characteristic image data; the RFID module then reads the unique tag ID on the bag, quickly associates it with the size characteristic data, and forms the final data packet for upload.

[0095] The visual recognition processing module inputs the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and simultaneously generates preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor;

[0096] It should be noted that the information printed on the woven bags is detected in real time through the YOLOv8s model. Combined with the pre-trained data set for transfer learning, it automatically matches the enterprise grading standards (such as Grade A ≥ 30cm, Grade B 20-30cm, and Grade C <20cm), and the results are written into the tamper-proof traceability chain of the blockchain database (Hyperledger Fabric).

[0097] Specifically, the application logic of the preliminary fusion data is:

[0098] Synchronously compare the acquisition timestamps corresponding to the target size classification information and the quick-freezing chamber temperature information. If the time difference is less than a preset time threshold, perform a structured combination of the target size classification information and the quick-freezing chamber temperature information to generate preliminary fusion data.

[0099] When the recognition credibility of the target size classification information or the data abnormality of the quick-freezing chamber temperature information exceeds the preset threshold, the multimodal interactive feedback module is triggered to issue an alarm and the manual mode is started to review the data to ensure the accuracy of data fusion.

[0100] The central control processing module synchronously inputs RFID tag information, preliminary fusion data, and fishing location data, and constructs a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules;

[0101] Specifically, the fishing location data is obtained by the GPS and Beidou positioning modules; the central control processing module integrates various sensor data, constructs a standard data packet, and uploads it to the cloud through the communication module.

[0102] The communication module uploads the full-link traceability data packet to the cloud data center in real time, and at the same time triggers the multimodal interactive feedback device to provide real-time feedback on the data binding status in the form of sound, light and touch.

[0103] Specifically, the communication module includes a built-in 4G / 5G cellular communication unit and a satellite communication unit, which are responsible for uploading real-time data to the cloud data center; the built-in encryption unit or the encryption unit within the main control unit completes the AES-256 encrypted transmission of data; and realizes full-link traceability records of squid fishing time, operating sea area, and responsible entities.

[0104] For example, the device integrates shipborne Wi-Fi (IEEE 802.11ax), 5G SA mobile network and satellite communication dual-redundant communication modules, built-in intelligent switching algorithm and AES-256-GCM data encryption tunnel, and automatically switches to 5G mode when the Wi-Fi signal is lower than -70dBm. It combines differential transmission and storage and forwarding mechanisms to ensure the complete upload of data in remote areas without network coverage; the data is finally stored on the Hyperledger Fabric blockchain platform to achieve tamper-proof data storage throughout the entire process.

[0105] The GPIO driver circuit built into the main control unit includes a three-color indicator light (LED driver), a buzzer sound prompt driver, and a motor or vibration motor driver circuit, providing three-dimensional feedback through sound, light, and touch. After the communication module uploads the information, the central control module drives the multimodal feedback module to provide real-time status feedback to personnel.

[0106] Specifically, the multimodal feedback module is a device feedback mechanism that integrates multiple sensory channels (such as vision, hearing, and touch). It is used to promptly and effectively convey the real-time status of the equipment operation or the results of data binding to the operator. In the squid fishing traceability operation: when the RFID tag and visual recognition automatic binding data are uploaded, the multimodal module immediately feedbacks green light and prompt sound; if the automatic mode cannot accurately read the size or RFID tag, the multimodal module feedback red light, strong sound and vibration to remind the staff to perform manual operation in time; when the manual operation mode is enabled, the yellow indicator light and appropriate sound prompts can remind personnel that the current system is in manual intervention mode.

[0107] Example 3

[0108] See also Figure 2As shown, this embodiment provides an intelligent method for tracing the quality of squid on board a ship, comprising the following steps:

[0109] S101: Capturing a first target image within a target area, reading RFID tag information of the first target, and marking the first target image and the corresponding RFID tag information as initial data;

[0110] S102: The automatic brim adjustment mechanism based on dynamic ambient light sensing dynamically adjusts the extension and retraction state of the brim, and adjusts the illumination angle and brightness in real time through the multi-light source lighting system to optimize the size feature image data of the first target image;

[0111] Specifically, the automatic adjustment mechanism of the brim based on dynamic ambient light sensing includes:

[0112] The ambient illumination value is collected in real time through the light sensor probe installed on the top of the device;

[0113] When the ambient illumination is lower than the ambient illumination threshold, the electric actuator is controlled to drive the brim to automatically retract;

[0114] When the ambient illumination is higher than the ambient illumination threshold, the electric actuator is controlled to drive the brim of the hat to automatically extend.

[0115] Based on the ambient light values ​​collected in real time by the light sensor 5, the light source intensity is dynamically adjusted and the light source angle is controlled to keep the image brightness constant within the camera's field of view.

[0116] S103: Inputting the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and generating preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor;

[0117] Specifically, the deep convolutional network model includes:

[0118] Construct a deep convolutional network with multi-scale convolution kernels, train it based on historical target size classification image data, and determine the convolution kernel parameters;

[0119] The optimized target size feature image is input into the deep convolutional network model, and the corresponding size classification label is output according to the preset size classification standard.

[0120] The method for determining the convolution kernel parameters includes:

[0121] Extract local features and overall features of historical target size images to generate multi-scale image feature data;

[0122] The multi-scale image feature data is input into the pre-trained deep network model and trained repeatedly until the model's accuracy in identifying target size reaches the preset standard.

[0123] S104: Synchronously input the RFID tag information, preliminary fusion data, and fishing location data, and construct a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules;

[0124] The data fusion algorithm and timestamp synchronization rules include:

[0125] The data obtained from RFID, visual recognition unit, temperature sensor, GPS and Beidou positioning module are respectively timestamped in a unified format;

[0126] Using the timestamp of fishing as a benchmark, the data from each sensor are time-aligned to form a time-synchronized squid traceability data packet.

[0127] The process of performing time sequence alignment on the sensor data includes:

[0128] Calculate the time difference between the collection time of RFID, visual, temperature and positioning data and the fishing reference time;

[0129] Data with a time difference less than the preset time threshold is regarded as valid synchronization data and integrated into the final traceability data packet.

[0130] S105: Upload the full-link traceability data packet to the cloud data center in real time, and trigger the multimodal interactive feedback device at the same time to provide real-time feedback on the data binding status in the form of sound, light and touch.

[0131] The multimodal interactive feedback device real-time feedback of the data binding status includes:

[0132] When the central control processing module confirms that the data packet is complete and valid, the control status indicator light turns green and a short sound prompt is triggered at the same time;

[0133] When the central control processing module detects missing or incorrect data, the control status indicator light turns red, accompanied by a warning sound and vibration feedback, to remind staff to review the data binding status.

[0134] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An intelligent method for tracing the quality of shipborne squid, characterized in that: include: S101: Capturing a first target image within a target area, reading RFID tag information of the first target, and marking the first target image and the corresponding RFID tag information as initial data; S102: The automatic brim adjustment mechanism based on dynamic ambient light sensing dynamically adjusts the extension and retraction state of the brim, and adjusts the illumination angle and brightness in real time through the multi-light source lighting system to optimize the size feature image data of the first target image; S103: Inputting the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and generating preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor; S104: Synchronously input the RFID tag information, preliminary fusion data, and fishing location data, and construct a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules; S105: Upload the full-link traceability data packet to the cloud data center in real time, and trigger the multimodal interactive feedback device at the same time to provide real-time feedback on the data binding status in the form of sound, light and touch.

2. The intelligent method for tracing the quality of squid on board a ship according to claim 1, characterized in that: The automatic hat brim adjustment mechanism based on dynamic ambient light sensing includes: The ambient illumination value is collected in real time through the light sensor probe installed on the top of the device; When the ambient illumination is lower than the ambient illumination threshold, the electric actuator is controlled to drive the brim to automatically retract; When the ambient illumination is higher than the ambient illumination threshold, the electric actuator is controlled to drive the brim of the hat to automatically extend; Based on the ambient light values ​​collected in real time by the light sensor 5, the light source intensity is dynamically adjusted and the light source angle is controlled to keep the image brightness constant within the camera's field of view.

3. The intelligent method for tracing the quality of squid on board a ship according to claim 2, characterized in that: The deep convolutional network model includes: Extract local and overall features of historical target size images to generate multi-scale image feature data; construct a deep convolutional network containing multi-scale convolution kernels and train it based on historical target size graded image data. Repeat the training until the model's accuracy in target size recognition reaches the preset standard, and determine the convolution kernel parameters; The optimized target size feature image is input into the deep convolutional network model, and the corresponding size classification label is output according to the preset size classification standard.

4. The intelligent method for tracing the quality of squid on board a ship according to claim 1, characterized in that: The data fusion algorithm and timestamp synchronization rules include: Calculate the time difference between the collection time of RFID, visual, temperature and positioning data and the fishing reference time; Data with a time difference less than the preset time threshold is considered valid synchronization data and is timestamped in a unified format. Using the timestamp of fishing as a benchmark, the data from each sensor are time-aligned to form a time-synchronized squid traceability data packet.

5. The intelligent method for tracing the quality of squid on board a ship according to claim 4 is characterized by: The multimodal interactive feedback device real-time feedback of the data binding status includes: When the central control processing module confirms that the data packet is complete and valid, the control status indicator light turns green and a short sound prompt is triggered at the same time; When the central control processing module detects missing or incorrect data, the control status indicator light turns red, accompanied by a warning sound and vibration feedback, to remind staff to review the data binding status.

6. A shipborne squid quality traceability intelligent system, based on the implementation of the shipborne squid quality traceability intelligent method according to any one of claims 1 to 5, characterized in that: The PCB, integrated into the terminal device, includes a data acquisition module, an environmental sensing module, a visual recognition processing module, a central control processing module, and a communication module. Each module is connected via wired and / or wireless communication, enabling real-time data binding and cloud synchronization throughout the entire squid catching, sorting, and packing process. A data acquisition module collects a first target image in the target area, reads the RFID tag information of the first target, and marks the first target image and the corresponding RFID tag information as initial data; The data acquisition module sends the initial data to the central control processing module; The environmental sensing module uses an automatic brim adjustment mechanism based on dynamic ambient light sensing to dynamically adjust the brim's extension and contraction state. It also uses a multi-light source lighting system to adjust the lighting angle and brightness in real time, optimizing the dimensional feature image data of the first target image. This dimensional feature image data is then fed into the visual recognition processing module. The visual recognition processing module inputs the optimized size feature image data into a deep convolutional network model to obtain current target size classification information; and simultaneously generates preliminary fusion data based on the target size classification information and the quick-freezing chamber temperature information obtained in real time from the temperature sensor; and sends the preliminary fusion data to the central control processing module; The central control processing module synchronously inputs RFID tag information, preliminary fusion data, and fishing location data, and constructs a full-link traceability data package based on the set data fusion algorithm and timestamp synchronization rules; The communication module uploads the full-link traceability data packet to the cloud data center in real time, and at the same time triggers the multimodal interactive feedback device to provide real-time feedback on the data binding status in the form of sound, light and touch.

7. The shipborne squid quality traceability intelligent system according to claim 6 is characterized by: The data acquisition module includes an automatic acquisition mode and a manual acquisition mode; In automatic acquisition mode, a high-resolution industrial camera equipped with a polarizing filter and an adaptive fill light module can identify the size markings on the squid woven bags in real time, read the label codes, and automatically associate them with GPS or Beidou positioning data to generate a unified data package. In manual collection mode, the squid size information is manually input through a physical button with anti-accidental touch logic, triggering the reading and binding of RFID tags. The bound data generates a data packet with a marked field, and the manual operation status is prompted through a multimodal feedback device.

8. The shipborne squid quality traceability intelligent system according to claim 7 is characterized by: The method for generating the preliminary fusion data is: The target size classification information is compared with the collection timestamp corresponding to the quick-freezing chamber temperature information, and structured data fusion is performed when the time difference between the two is less than the preset threshold; when the credibility of the size classification information or the temperature information abnormality exceeds the set threshold, the multimodal feedback device alarm is automatically triggered and the manual review mode is started.

9. An intelligent method for tracing the quality of squid on board a ship, for integrating and fixing the intelligent system for tracing the quality of squid on board a ship according to any one of claims 6 to 8, characterized in that: The invention comprises a conveyor rolling track (7) fixed on a height-adjustable bracket (8), wherein an anti-corrosion shell is provided on the side of the height-adjustable bracket (8) where the conveyor rolling track (7) is located, and the anti-corrosion shell comprises: An RFID identification antenna (1) is used to cooperate with an RFID reader to transmit a radio frequency signal to activate a passive RFID tag in the squid woven bag and receive unique coding information returned by the tag; Display screen (2), showing the equipment running status, data binding status, real-time temperature, size grade display, and providing a manual interaction interface in manual mode; Button (3), used for the operator to quickly select the squid size class by pressing a button in manual mode; Status indicator (4), real-time display of device operating status and data binding results: The light sensor (5) monitors the light intensity of the on-site environment in real time; the light intensity is sent to the light sensing detection module to control the automatic extension and retraction of the brim of the device and the automatic adjustment of the brightness of multiple light sources; The positioning antenna (6) realizes GPS / Beidou satellite positioning, obtains the squid fishing location, and transmits the squid fishing location data.

10. The intelligent method for tracing the quality of squid carried on board a ship according to claim 9, characterized in that: The RFID identification antenna (1) adopts a "three-section" identification antenna, and RFID radio frequency antenna segments are respectively installed in the first two sections of the three-section structure; the two RFID radio frequency antenna segments form an angle of about 30 degrees in the spatial layout, presenting a front-to-back staggered and angled setting, forming a spatially differentiated identification channel.

Images