A smart feeding control system and method for large-scale cultivation of Noctiluca scintillans

Abstract

This invention discloses an intelligent feeding control system and method for large-scale cultivation of *Noctiluca scintillans*, belonging to the field of algae cultivation technology. The system includes an image recognition and counting module, an online concentration monitoring module, a feeding decision control module, and an automatic feeding execution module. These modules are interconnected to form a complete feeding control architecture. A growth feedback correction module can be optionally added to further enhance adaptability. The image recognition and counting module acquires the number of *Noctiluca scintillans*, the online concentration monitoring module obtains the concentration of the food algae through optical detection, the feeding decision control module calculates the feeding volume according to the formula V=k×N / C, and can dynamically adjust the feeding ratio coefficient k according to the cultivation stage, and the automatic feeding execution module performs quantitative feeding. This invention achieves precise and automated control of intelligent feeding of *Noctiluca scintillans* and real-time monitoring and regulation of the culture medium and environment, adapting to the needs of large-scale cultivation, reducing food waste and water quality fluctuations, improving the proliferation efficiency of *Noctiluca scintillans*, and has strong practicality.

Description

Technical Field

[0001] This invention belongs to the field of algae cultivation technology, specifically relating to an intelligent feeding control system and method for large-scale cultivation of Noctiluca scintillans. Background Technology

[0002] Noctiluca scintillans is a single-celled planktonic organism belonging to the family Noctilucidae in the phylum Dinophyta, widely distributed in coastal waters worldwide. Noctiluca scintillans mainly includes green Noctiluca scintillans, which lives in symbiosis with photosynthetic microalgae, and red Noctiluca scintillans, which is purely heterotrophic. Red Noctiluca scintillans is the most common along the coast of my country and is also an important species contributing to the "blue tears" phenomenon along the coast from Pingtan to Xiamen in Fujian. This thallus is nearly sac-shaped, relatively large, and visible to the naked eye. It lacks photosynthetic capacity and obtains nutrients by feeding on phytoplankton and organic particles. When stimulated, it emits blue-green fluorescence, forming a striking bioluminescent spectacle when aggregated.

[0003] Noctiluca scintillans possesses value in cultural tourism display, ecological research, and artificial cultivation and development. On the one hand, "blue tears" have high ornamental and scientific display potential, and can be used for cultural tourism landscape creation and live exhibitions; on the other hand, Noctiluca scintillans is also a typical red tide-related organism, and its massive reproduction can affect the nearshore ecological environment and fishery production. Therefore, conducting research on the growth patterns and artificial cultivation of Noctiluca scintillans is significant for the development of ornamental resources, the study of red tide mechanisms, and ecological control.

[0004] While existing technologies disclose automated cultivation devices, breeding equipment, and rapid proliferation or mixed cultivation methods for *Noctiluca scintillans*, they mostly rely on preset feeding times, fixed cultivation conditions, fixed initial concentrations of food algae, or predetermined cultivation regimes for control. They lack closed-loop intelligent feeding control schemes based on real-time *Noctiluca scintillans* quantity sensing and online monitoring of food algae concentration, thus lacking intelligent detection, analysis, feedback, and regulation functions. Meanwhile, *Noctiluca scintillans* cultivation generally suffers from long cultivation cycles, significant influence of food supply on proliferation efficiency, amplified feeding errors as cultivation scale increases, and decreased cultivation stability. Due to the thin cell walls and sensitivity to disturbance of *Noctiluca scintillans*, conventional empirical feeding methods are insufficient to simultaneously achieve high-density cultivation and stable proliferation.

[0005] Therefore, it is necessary to establish an intelligent control system for the cultivation process of Noctiluca scintillans. By sensing the number of Noctiluca scintillans and the concentration of food algae in real time, and combining it with a pre-constructed feeding relationship model, the system can dynamically calculate and execute feeding operations, thereby improving the cultivation efficiency of Noctiluca scintillans, reducing the intensity of artificial intervention, and providing technical support for the large-scale cultivation and related applications of Noctiluca scintillans. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent feeding control system and method for the large-scale cultivation of Noctiluca scintillans. This invention breaks through the limitations of the prior art, which relies on preset feeding regimes and lacks real-time sensing and closed-loop control. It enables online monitoring of Noctiluca scintillans quantity and feed concentration, and dynamic and precise control of feeding amount, thereby improving cultivation stability and efficiency and providing technical support for the large-scale cultivation of Noctiluca scintillans.

[0007] This invention provides an intelligent feeding control system for the large-scale cultivation of Noctiluca scintillans, comprising:

[0008] Image recognition and counting module, online concentration monitoring module, feeding decision control module, and automatic feeding execution module;

[0009] The feeding decision control module is communicatively connected to the image recognition and counting module, the online concentration monitoring module, and the automatic feeding execution module, respectively.

[0010] The image recognition and counting module is used to obtain the number of Noctiluca scintillans;

[0011] The online concentration monitoring module is used to obtain the concentration of food algae;

[0012] The feeding decision control module is used to calculate the feeding volume based on the number of Noctiluca scintillans, the concentration of food algae, and the feeding ratio coefficient.

[0013] The automatic feeding execution module is used to perform quantitative feeding under the control of the feeding decision control module.

[0014] Preferably, the image recognition and counting module performs individual identification and counting of the acquired Noctiluca scintillans images through image processing analysis; the online concentration monitoring module uses optical detection and converts the detection signal into the concentration of food algae cells; the feeding decision control module calculates the feeding volume according to the formula V=k×N / C, where N is the number of Noctiluca scintillans, C is the concentration of food algae, and k is the feeding ratio coefficient. The feeding decision control module dynamically adjusts the feeding ratio coefficient k according to the cultivation stage.

[0015] Optionally, the system further includes a growth feedback correction module, which is connected to the feeding decision control module and corrects the feeding ratio coefficient based on the growth data of Noctiluca scintillans.

[0016] This invention also provides an intelligent feeding control method for large-scale cultivation of Noctiluca scintillans, comprising:

[0017] The concentration of bait algae was obtained through online monitoring;

[0018] The number of Noctiluca scintillans was obtained through image recognition;

[0019] The feeding volume is calculated based on the number of Noctiluca scintillans, the concentration of bait algae, and the feeding ratio coefficient.

[0020] Perform quantitative feeding according to the stated feeding volume.

[0021] Furthermore, the method also includes: adjusting the feeding ratio coefficient based on the growth data of Noctiluca scintillans.

[0022] Furthermore, the method dynamically adjusts the feeding ratio coefficient according to the cultivation stage.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] 1. Achieve precise and intelligent feeding of Noctiluca scintillans.

[0025] By using image recognition to count Noctiluca scintillans in real time and monitor the concentration of food algae online, combined with a multi-parameter dynamic model to calculate the feeding amount, the traditional fixed-rhythm feeding mode is broken away from, significantly improving the feeding accuracy.

[0026] 2. Improve the stability and success rate of Noctiluca scintillans proliferation.

[0027] Based on real-time growth status and closed-loop control of feed concentration, insufficient or excessive feeding is avoided, the impact of environmental disturbance on Noctiluca scintillans is reduced, and the stability and proliferation efficiency of high-density culture are greatly improved.

[0028] 3. Reduce reliance on manual labor and minimize human error.

[0029] It eliminates the need for frequent offline sampling, microscopic counting, and experience-based judgment, enabling fully automated and intelligent operation, significantly reducing human intervention, and making it suitable for large-scale continuous culture.

[0030] 4. The cultivation process can be dynamically and adaptively adjusted.

[0031] The system can correct model parameters based on feedback from actual growth data of Noctiluca scintillans, and automatically adjust feeding strategies at different cultivation stages, making it more adaptable and the cultivation system more stable.

[0032] 5. More suitable for large-scale propagation and cultural tourism live exhibition scenarios

[0033] This system can operate stably for a long time without being limited by fixed formulas or fixed initial concentrations, and can meet the engineering needs of large-scale propagation, continuous cultivation and live landscape display of Noctiluca scintillans. Attached Figure Description

[0034] Figure 1 Schematic diagram of the intelligent feeding control system for Noctiluca scintillans (Example 1);

[0035] Figure 2 Schematic diagram of static image acquisition of Noctiluca scintillans;

[0036] Figure 3 Schematic diagram of the image recognition and counting process for Noctiluca scintillans; Figure 3 A in the image is the original image; Figure 3 B in the image is a binarized segmented image;

[0037] Figure 4 A graph showing the results of connected component analysis and counting;

[0038] Figure 5 This is a mean growth curve of Platycladus orientalis, reflecting the density change trend of Platycladus orientalis during the cultivation period;

[0039] Figure 6 A graph showing the relationship between cell concentration and chlorophyll a content under different dilution gradients of Platycodon grandiflorus;

[0040] Figure 7 The graph shows the relationship between the absorbance of different concentrations of Platycodon grandiflorus algal solutions at 680 nm and cell concentration.

[0041] Figure 8 Growth curves of Noctiluca scintillans under different feeding volume gradients for Platycodon grandiflorus;

[0042] Figure 9 Growth curves of Noctiluca scintillans under different concentration gradients of Platycladus orientalis;

[0043] Figure 10 Schematic diagram of the intelligent feeding control system for Noctiluca scintillans (Example 2);

[0044] Figure 11 Flowchart of the intelligent feeding control method for Noctiluca scintillans. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Example 1

[0047] This embodiment provides an intelligent feeding control system for the large-scale cultivation of Noctiluca scintillans, such as... Figure 1 As shown, it includes: an image recognition and counting module, an online concentration monitoring module, a feeding decision control module, and an automatic feeding execution module; the feeding decision control module is communicatively connected to the image recognition and counting module, the online concentration monitoring module, and the automatic feeding execution module, respectively.

[0048] 1. Image recognition and counting module

[0049] To achieve quantity sensing during the cultivation of *Noctiluca scintillans*, this invention includes an image recognition and counting module. This module is used to acquire static images of *Noctiluca scintillans* in the cultivation tank, and completes target identification and quantity counting of *Noctiluca scintillans* through image processing. It provides input data on the quantity of *Noctiluca scintillans* for the subsequent feeding decision control module, and is the core sensing unit for achieving precise feeding.

[0050] (1) Module composition and function

[0051] The image recognition and counting module includes an image acquisition unit, an imaging assistance unit, an image processing unit, a counting output unit, and a control connection unit. These units work together to complete the entire process from image acquisition to data output.

[0052] Image acquisition unit: used to acquire static images of the Noctiluca scintillans culture pond, providing raw data for subsequent identification and counting;

[0053] Imaging assistance unit: Used to provide stable shooting conditions, including fixed brackets, background boards, supplementary light sources or light-blocking structures, to ensure stable image quality;

[0054] Image processing unit: Used for preprocessing, target segmentation and quantity statistics of acquired images, and is the core algorithm unit of the module;

[0055] Counting output unit: Used to output the number of Noctiluca scintillans, density per unit field of view, or total amount of culture pond after conversion, and to complete the standardized output of counting results;

[0056] Control connection unit: used to send the counting results to the feeding decision control module, participate in the subsequent feeding volume calculation, and realize data interaction between modules.

[0057] (2) Image acquisition method Figure 2 )

[0058] The image acquisition unit preferably employs a high-definition industrial camera, high-definition webcam, or image acquisition device with macro shooting capabilities. It is installed above, to the side, or diagonally above the *Noctiluca scintillans* culture container to acquire static image information of the *Noctiluca scintillans* in the culture medium. To ensure imaging consistency, the image acquisition unit is used in conjunction with a fixed support to maintain stability in the shooting angle, shooting distance, and focal plane.

[0059] To improve target imaging quality, the imaging auxiliary unit can be equipped with a background plate, a ring-shaped supplementary light source, a backlighting device, or a light-shielding structure to reduce the impact of reflections from the culture container wall, liquid surface glare, bubble interference, and background noise on the recognition results. Image acquisition adopts a timed photography method, acquiring images of Noctiluca scintillans at preset time points and transmitting the images to the image processing unit.

[0060] (3) Image recognition and counting process

[0061] After receiving the image data acquired by the image acquisition unit, the image processing unit executes the following processing steps in sequence to complete the automatic identification and counting of Noctiluca scintillans:

[0062] Image preprocessing: The original image is cropped, converted to grayscale, background corrected, Gaussian denoising and median smoothing are performed in sequence to eliminate interference such as stray light, bubbles and impurities in the image and improve image clarity;

[0063] Binarization processing: The binarization threshold is automatically adjusted based on the grayscale distribution characteristics of the image to accurately separate the Noctiluca scintillans target area from the background area, thereby obtaining the target candidate area. Figure 3 );

[0064] Connectivity analysis: Independent target regions in the binary image are identified and marked. Simultaneously, based on the morphological characteristics of *Noctiluca scintillans*, screening thresholds for area, roundness, and aspect ratio are set to exclude non-*Noctiluca scintillans* targets such as bubbles, impurities, and wall reflections. Figure 4 );

[0065] Overlapping target segmentation: For Noctiluca scintillans targets that are close to each other or partially overlap during high-density cultivation, a segmentation method of "circular contour fitting + pixel area segmentation" is adopted based on their near-spherical sac-like morphological characteristics to reduce the missed count rate.

[0066] Target identification and counting: Identify individual Noctiluca scintillans algae based on their outline size, morphological features, and area range, and output the number of Noctiluca scintillans algae within the field of view;

[0067] Quantity conversion: Based on the field of view area, liquid layer thickness, imaging area scale, or preset conversion parameters, the counting results within the field of view are converted into the density of Noctiluca scintillans per unit volume or the total number of Noctiluca scintillans in the culture tank.

[0068] The above process enables automatic detection and non-contact counting of Noctiluca scintillans, significantly reducing the time cost and subjective errors associated with manual counting.

[0069] (4) Identify constraints and error control

[0070] Due to potential interference factors such as air bubbles, impurities, wall reflections, and the proximity of high-density targets during the cultivation of Noctiluca scintillans, the image processing unit sets the following constraints to improve counting accuracy:

[0071] Set screening thresholds for the area range, roundness, or morphological characteristics of connected regions to exclude obvious non-Noctiluca scintillans targets and reduce the false detection rate.

[0072] For targets that are close to each other or partially overlap, perform contour correction or region splitting to reduce the omission rate;

[0073] Filtering out abnormal bright spots, reflective points, or random noise further improves recognition accuracy;

[0074] By adjusting the shooting position, lighting conditions, and image processing threshold, the consistency of counting between different culture batches can be improved, thus ensuring data stability.

[0075] 2. Online Concentration Monitoring Module

[0076] To achieve real-time and accurate sensing of the concentration of bait algae (flat algae), this invention sets up an online concentration monitoring module, which is integrated into the flat algae cultivation photoreactor. This module is used to monitor the concentration of the flat algae solution in real time and input the concentration data into the feeding decision control module, providing core parameters for feeding volume calculation.

[0077] (1) Module composition and detection method

[0078] The core of the online concentration monitoring module is the algal solution density detection unit, which can use OD... 680 Absorbance detection probe, chlorophyll fluorescence detection probe, scattering spectroscopy detection device, or other online detection devices suitable for characterizing the concentration of Platycladus orientalis. Among them, OD... 680 The detection method is the preferred solution of this invention because it is fast, easy to operate, sensitive to changes in the concentration of chlorophyll-containing algal solution, and suitable for online continuous monitoring.

[0079] The preferred installation location for the detection unit is the uniform mixing area in the middle of the photoreactor body, the liquid circulation loop, or the bypass flow detection pool, in order to reduce the impact of algal sedimentation, bubble adhesion, and local uneven concentration on the detection results and ensure the stability and accuracy of the detection data.

[0080] The density change trend of Platycodon grandiflorus during the culture period is as follows: Figure 5 As shown, this visually reflects the complete growth pattern of Platycladus orientalis from inoculation to the stationary phase, providing a basis for the design and calibration of concentration detection ranges.

[0081] (2) Calibration and concentration conversion model

[0082] To achieve precise conversion of optical signals to cell concentration, a calibration process must be completed before the system can be run.

[0083] Zero-point calibration was performed using blank culture medium to eliminate background interference;

[0084] Multiple Platycodon grandiflorus standard samples with known cell concentration gradients were prepared, and their optical signals and actual cell concentrations were measured simultaneously (obtained by methods such as microscopic counting).

[0085] A conversion relationship between optical signals and Platycodon cell concentration was established through linear fitting.

[0086] Experiments have verified that the concentration of Platycladus orientalis cells is related to chlorophyll a content and OD. 680 The absorbance values ​​all showed a good linear correlation, and the corresponding relationships are as follows: Figure 6 , Figure 7 As shown.

[0087] based on Figure 7 The calibration results showed that, within the suitable concentration range for Platycladus orientalis culture, the cell concentration of Platycladus orientalis was related to OD. 680 The values ​​show a good linear relationship, and the fitting equation is:

[0088] OD680 = 7.659 × 10 -7 ×C+0.005563

[0089] In the formula, C represents the concentration of Platychophyton floccosum cells, measured in cells / mL.

[0090] Convert the above equation into a concentration calculation formula:

[0091] C=(OD680-0.005563) / (7.659×10 -7 )

[0092] In addition to OD 680 In addition to detection, this module can also use chlorophyll fluorescence or other spectral detection methods to establish a conversion model between detection signals and cell concentration, adapting to different application scenarios.

[0093] (3) Actual operation and data output

[0094] During actual system operation, the online concentration monitoring module collects OD data in real time. 680 The system uses values ​​or other optical detection signals to calculate the cell concentration of the algae solution in the current photoreactor through a pre-calibrated conversion model, and outputs the concentration data stably to the feeding decision control module to provide a basis for calculating the volume of algae solution to be added.

[0095] To ensure long-term detection accuracy, the system can be configured with probe cleaning, periodic recalibration, temperature compensation, and manual verification and correction mechanisms to correct detection deviations caused by factors such as probe drift and surface contamination, providing continuous and accurate concentration data for feeding decisions.

[0096] 3. Feeding Decision Control Module

[0097] To achieve precise feeding control in the cultivation of Noctiluca scintillans, this invention sets up a feeding decision control module, which serves as the control core of the system. This module constructs a feeding parameter model through preliminary feeding relationship calibration experiments, and then dynamically calculates the feeding volume of Noctiluca scintillans by combining real-time monitoring data on the number of Noctiluca scintillans and the concentration of Platycladus orientalis, ultimately achieving adaptive and refined feeding control.

[0098] (1) Determination of the feeding relationship between Noctiluca scintillans and Platycodon grandiflorus

[0099] To establish a feeding decision parameter system suitable for Noctiluca scintillans cultivation, this invention first conducts a calibration experiment on the feeding relationship between Noctiluca scintillans and Platycladus orientalis to obtain empirical parameter relationships between feed supply levels and Noctiluca scintillans proliferation response, providing boundary and coefficient basis for subsequent calculation models.

[0100] ①Calibration Experiment Design

[0101] Two types of gradient experiments were used to determine the feeding relationship:

[0102] Experiments with different feeding volume gradients: Under the condition that the initial number of Noctiluca scintillans is the same, multiple groups of feeding volumes of Noctiluca scintillans were set up to compare the trend of Noctiluca scintillans number changes during the culture period;

[0103] Experiments with different concentration gradients of Noctiluca scintillans: Under the condition that the initial number of Noctiluca scintillans is the same, multiple groups of Noctiluca scintillans concentration levels were set up to compare the differences in the proliferation response of Noctiluca scintillans.

[0104] The experiment was conducted under fixed culture volume, temperature, salinity, and light conditions. The number of *Noctiluca scintillans* and the corresponding *Platycodon grandiflorus* concentration / feeding level were recorded at preset time intervals. Optimal conditions: feeding was performed every two days, the experimental period was 7–10 days, and each group had 3 parallel replicates to reduce experimental error and improve the stability of parameter extraction.

[0105] ② Evaluation indicators for the growth response of Noctiluca scintillans

[0106] Based on the time series data for each group, the following growth indicators were calculated:

[0107] Stage growth slope: reflects the proliferation rate of Noctiluca scintillans per unit time;

[0108] Endpoint number of individuals: reflects the overall proliferation level during the culture period;

[0109] Multiplication factor: reflects the degree of growth relative to the initial quantity;

[0110] Growth trend stability: reflects whether growth continues, or whether there is a plateau or stagnation.

[0111] By comparing multiple indicators, the growth effect of Noctiluca scintillans under different feeding levels was determined, and the quantitative relationship between the number of Noctiluca scintillans, the supply level of Platycladus orientalis, and the proliferation response was extracted.

[0112] ③ Calibration Experiment Results and Trend Analysis

[0113] Under different feeding volumes of *Platycodon grandiflorus*, the overall number of *Noctiluca scintillans* individuals showed an increasing trend, and the proliferation rate significantly increased with increasing feeding levels. The high feeding volume group had a higher number of individuals and a higher growth slope in the middle and later stages of cultivation, indicating that increasing the food supply is beneficial to the continuous proliferation of *Noctiluca scintillans*. The results are shown in [Figure 1]. Figure 8 With increasing feeding volume, the growth rate and the number of individuals at the endpoint of Noctiluca scintillans significantly increased.

[0114] The growth curves of *Noctiluca scintillans* differed significantly under different *Platycodon grandiflorus* concentrations. The higher concentration group showed faster proliferation and a higher final number of individuals, indicating that *Noctiluca scintillans* proliferation has a significant response to the concentration of its food algae. The results are shown in [Figure number missing]. Figure 9 Within the experimental range, the higher concentration group of Noctiluca scintillans showed a better trend of continuous growth.

[0115] Overall trend analysis indicates a stable positive correlation between the proliferation effect of *Noctiluca scintillans* and the supply level of *Platycodon grandiflorus*: growth is slow when the feeding level is low; and the growth slope and the final number of individuals increase significantly after the feeding level is increased. Therefore, the optimal feeding relationship should be expressed as an interval range rather than a single fixed value.

[0116] ④ Recommended feeding ratio range

[0117] Based on the parameter data of the group with better overall growth effect, combined with the stage growth slope, the number of individuals at the endpoint, and growth stability, the range of feeding ratios was determined:

[0118] Recommended range: Flat algae : Noctiluca scintillans = (1000~1400) : 1

[0119] Preferred range: Flat algae : Noctiluca scintillans = (1100~1300) : 1

[0120] This ratio represents the target number of *Platycodon* cells fed per unit quantity of *Noctiluca scintillans*. Within this range, *Noctiluca scintillans* can maintain good continuous growth and high proliferation efficiency; below this range, growth is insufficient, and above this range, economic efficiency and stability decrease. Therefore, the above range is the preferred control range.

[0121] (2) Feeding calculation model

[0122] To apply the calibration results to actual culture control, an empirical calculation model based on the feeding coefficient is established, with the following parameters defined:

[0123] N: Current number of Noctiluca scintillans;

[0124] C: Algal concentration of Platycladus orientalis solution before feeding (cells / mL);

[0125] k: Recommended feeding coefficient (number of target flat algae cells corresponding to one unit of Noctiluca scintillans);

[0126] F: Total number of target flat algae cells fed;

[0127] V: Volume of algal solution to be added.

[0128] ① Target total number of cells fed

[0129] F=k×N

[0130] Among them, k is preferably taken as 1000~1400, and further preferably 1100~1300.

[0131] ② Volume of algae solution to be added

[0132] V=k×N / C

[0133] Given the number of Noctiluca scintillans (N) and the concentration of Platycladus orientalis (C), the required volume of feed can be directly calculated.

[0134] (3) Practical application of the model and phased control

[0135] In the actual cultivation process, the feeding decision control module executes the following steps sequentially: The real-time Noctiluca scintillans count N is obtained through the image recognition and counting module; the real-time Platycladus orientalis concentration C is obtained through the online concentration monitoring module; an appropriate feeding coefficient k is selected based on the cultivation stage; the target feeding volume is calculated according to the formula V=k×N / C; and the volume command is sent to the automatic feeding execution module to complete the quantitative addition. Simultaneously, this module adopts a phased differentiated control strategy: during the adaptation stage, k is set close to the lower limit of the recommended range to reduce cultivation stress; during the rapid proliferation stage, k is set to the median or upper limit to ensure the growth rate; and during the stable maintenance stage, k is dynamically fine-tuned based on growth feedback.

[0136] (4) Dynamic decision-making and parameter correction

[0137] The feeding decision control module performs dynamic calculations based on the following parameters: real-time quantity of Noctiluca scintillans, historical growth slope, real-time concentration of Platycladus orientalis, historical feeding amount, feeding interval, culture stage, and parameter boundaries and appropriate feeding ratio determined by calibration experiments.

[0138] The feeding ratio coefficient k is a dynamically adjustable parameter, not a fixed value, and can be adjusted in real time according to the cultivation stage, the growth effect of the previous cycle, the remaining feed status, and the preset growth target.

[0139] The core feature of this module is that it does not feed according to a fixed time sequence. Instead, it dynamically calculates the feeding amount based on real-time monitoring data and historical growth response, and continuously optimizes parameters during continuous cultivation to achieve precise and adaptive feed supply control.

[0140] 4. Automatic feeding execution module

[0141] The automatic feeding execution module mainly includes a storage container, infusion pipeline, metering pump, control valve, and optional flow detection element. The feeding decision control module controls the running time, output speed, or valve opening time of the metering pump according to the calculated target feeding volume, so as to accurately deliver the corresponding volume of algal solution from the photoreactor or storage unit to the Noctiluca scintillans culture system, thereby realizing automated and quantitative feed feeding.

[0142] This module can use a peristaltic pump, syringe pump, metering pump, or a combination thereof as the execution unit. It is preferably paired with a flow detection element, valve control unit, and cleaning branch to further improve feeding accuracy and long-term operational stability. For the high-precision feeding requirements of *Noctiluca scintillans* cultivation, the module supports small-volume, multi-stage, or timed addition. Considering that *Noctiluca scintillans* is sensitive to fluid disturbances, a slow-flow-rate continuous dripping mode is used for multi-stage addition to avoid drastic local environmental fluctuations caused by a single large-volume addition, ensuring stable cultivation conditions.

[0143] To prevent algal sedimentation and aggregation, pipeline blockage, and pipeline contamination caused by long-term operation, the infusion pipeline can be equipped with a filtration structure, flushing branches, or a periodic automatic cleaning program. Simultaneously, the system is equipped with idling monitoring, blockage alarms, low-liquidity alarms, and abnormal shutdown protection mechanisms. When abnormal probe signals, image recognition counting module failure, pump malfunction, or communication interruption are detected, the automatic feeding module can automatically pause feeding, enter a preset safe operating mode, and issue an alarm, significantly improving the safety and reliability of system operation.

[0144] Example 2

[0145] This embodiment provides an intelligent feeding control system for the large-scale cultivation of Noctiluca scintillans, and its system schematic diagram is shown below. Figure 10 As shown in the figure. This embodiment adds a growth feedback correction module to the existing embodiment 1, forming a complete closed-loop architecture of "perception-decision-execution-feedback-optimization". By dynamically correcting the feeding parameters through growth status feedback, the long-term adaptive capability and cultivation stability of the system are improved. In this embodiment, the hardware composition and working principle of the image recognition counting module, the online concentration monitoring module, and the automatic feeding execution module are the same as those in embodiment 1, and will not be repeated here.

[0146] 1. System Overall Architecture

[0147] The system in this embodiment includes: an image recognition and counting module, an online concentration monitoring module, a feeding decision control module, an automatic feeding execution module, and a growth feedback correction module. The communication connections between the modules are as follows:

[0148] The image recognition and counting module and the online concentration monitoring module output their respective detection data to the feeding decision control module. The image recognition and counting module also outputs continuous Noctiluca scintillans quantity data to the growth feedback correction module. The growth feedback correction module sends the parameter correction results to the feeding decision control module. The feeding decision control module outputs the generated feeding instructions to the automatic feeding execution module, realizing the coordinated linkage between the modules.

[0149] 2. Growth Feedback Correction Module

[0150] This module is used to dynamically adjust the feeding strategy based on the actual growth effect of Noctiluca scintillans, so as to achieve closed-loop optimization of feeding parameters.

[0151] After one or more feedings are completed, the growth feedback correction module compares the actual growth results with the preset growth target based on the subsequently collected data on the changes in the number of Noctiluca scintillans, and corrects the feeding ratio coefficient or model parameters to gradually improve the adaptability and stability of feeding control.

[0152] (1) Deviation Judgment Criteria

[0153] Deviation judgment criteria may include:

[0154] The difference between the actual growth slope and the target growth slope;

[0155] The number of animals has increased at a low rate for several consecutive feeding cycles;

[0156] The concentration of residual feed is too high after feeding;

[0157] The number of Noctiluca scintillans has decreased or its growth has stagnated.

[0158] (2) Parameter correction logic

[0159] When the actual growth is lower than the preset target and below the threshold range, the growth feedback correction module can increase the feeding ratio coefficient k by a preset step size to improve the feed supply level and ensure the proliferation needs of Noctiluca scintillans. When the remaining feed is detected to be too high or the growth enters a plateau period, the growth feedback correction module can decrease the feeding ratio coefficient k by a preset step size to avoid feed waste and water quality fluctuations and maintain the stability of the culture system.

[0160] (3) Model self-learning optimization

[0161] During continuous system operation, the growth feedback correction module can use multi-cycle culture data to update the model database, thereby gradually optimizing feeding parameters and improving prediction accuracy and control performance in subsequent culture processes. This feedback correction process constitutes an important part of the system's closed-loop control, enabling feeding parameters to adaptively adjust with changes in culture status, adapting to the personalized needs of different culture batches and different growth stages without manual intervention.

[0162] 3. Feeding Decision Control Module

[0163] This module serves as the core of the system control. Based on Example 1, it adds the input of the growth feedback correction module and integrates multi-source data to complete the decision-making process. The specific steps are as follows:

[0164] Receive the real-time quantity N from the image recognition and counting module and the real-time algal concentration C from the online concentration monitoring module;

[0165] The initial feeding coefficient k was determined by combining the initial calibration model and the cultivation stage;

[0166] Receive correction parameters from the growth feedback correction module and update the k value;

[0167] The target feeding volume is calculated using the formula V=k×N / C and sent to the automatic feeding execution module;

[0168] Store the data from this feeding to provide historical data support for the growth feedback correction module.

[0169] 4. System closed-loop workflow

[0170] Sensing phase: The image recognition and counting module completes the counting of Noctiluca scintillans and outputs N, while the online concentration monitoring module detects the concentration of Platycladus orientalis and outputs C;

[0171] Initial Decision: The feeding decision control module generates initial feeding instructions based on the initial model and stage information;

[0172] Execution phase: The automatic feeding module completes the quantitative feeding;

[0173] Data accumulation: The image recognition and counting module continuously collects quantity data, and the feeding decision control module stores the operation records;

[0174] Feedback optimization: The growth feedback correction module assesses the growth status, determines the deviation, corrects the parameters, and updates the feeding decision model;

[0175] Loop optimization: The feeding decision control module uses the optimized parameters for the next feeding, forming a closed-loop adaptive control.

[0176] Example 3

[0177] This embodiment provides an intelligent feeding control method for large-scale cultivation of Noctiluca scintillans, the flowchart of which is shown below. Figure 11 As shown, the specific implementation process of this method is as follows:

[0178] Step S1: Perform online optical detection on the algal solution in the algal culture unit. Through spectral analysis or turbidity conversion, convert the detection signal into algal cell concentration C, and obtain the actual concentration data of algal in real time to provide basic input for feeding calculation.

[0179] Step S2: Acquire static images of *Noctiluca scintillans* in the culture tank. Using algorithms such as image preprocessing, binarization segmentation, and connected component analysis, identify and count individual algae to obtain the number of algae, N, achieving real-time, non-contact monitoring of *Noctiluca scintillans* biomass. This step and Step S1 are parallel and independent sensing processes; there is no mandatory order, and they can be executed synchronously or in any order without affecting subsequent calculation results.

[0180] Step S3: Using the number of Noctiluca scintillans N and the concentration of food algae C as inputs, and combining the preset feeding ratio coefficient k, calculate the precise feeding volume according to the formula V=k×N / C; where the feeding ratio coefficient k can be dynamically adjusted according to different cultivation stages of Noctiluca scintillans (such as logarithmic growth phase, plateau phase) to adapt to the nutritional needs of different growth stages.

[0181] Step S4: Based on the calculated feeding volume, control the automatic feeding execution module to quantitatively transport the food algae from the culture unit to the Noctiluca scintillans culture tank, completing one intelligent feeding operation and realizing precise, closed-loop feeding control for Noctiluca scintillans cultivation.

[0182] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. An intelligent feeding control system for large-scale cultivation of Noctiluca scintillans, characterized in that, include: Image recognition and counting module, online concentration monitoring module, feeding decision control module, and automatic feeding execution module; The feeding decision control module is communicatively connected to the image recognition and counting module, the online concentration monitoring module, and the automatic feeding execution module, respectively. The image recognition and counting module is used to obtain the number of Noctiluca scintillans; The online concentration monitoring module is used to obtain the concentration of food algae; The feeding decision control module is used to calculate the feeding volume based on the number of Noctiluca scintillans, the concentration of food algae, and the feeding ratio coefficient; The automatic feeding execution module is used to perform quantitative feeding under the control of the feeding decision control module.

2. The system according to claim 1, characterized in that, The image recognition and counting module performs individual identification and counting of the acquired Noctiluca scintillans images through image processing and analysis.

3. The system according to claim 1, characterized in that, The online concentration monitoring module uses optical detection and converts the detection signal into the concentration of food algae cells.

4. The system according to claim 1, characterized in that, The feeding decision control module calculates the feeding volume according to the formula V=k×N / C, where N is the number of Noctiluca scintillans, C is the concentration of bait algae, and k is the feeding ratio coefficient.

5. The system according to claim 4, characterized in that, The feeding decision control module dynamically adjusts the feeding ratio coefficient k according to the cultivation stage.

6. The system according to claim 1, characterized in that, The system also includes a growth feedback correction module, which is connected to the feeding decision control module and corrects the feeding ratio coefficient based on the growth data of Noctiluca scintillans.

7. A method for intelligent feeding control in the large-scale cultivation of Noctiluca scintillans, characterized in that, include: The concentration of bait algae was obtained through online monitoring; The number of Noctiluca scintillans was obtained through image recognition; The feeding volume is calculated based on the number of Noctiluca scintillans, the concentration of bait algae, and the feeding ratio coefficient. Perform quantitative feeding according to the stated feeding volume.

8. The method according to claim 7, characterized in that, The method further includes: adjusting the feeding ratio coefficient based on the growth data of Noctiluca scintillans.

9. The method according to claim 8, characterized in that, The method dynamically adjusts the feeding ratio coefficient according to the cultivation stage.

Images