Intelligent organic synthesis reaction regulation and control method and system based on visual recognition

Abstract

The invention relates to the technical field of intelligent regulation and control, in particular to an organic synthesis reaction intelligent regulation and control method and system based on visual identification, and the method comprises the steps: obtaining a first image of a raw material dropwise adding funnel and a second image of a product collecting port; identifying the liquid level change of the funnel based on the first image, and determining the raw material dripping speed based on the liquid level change of the funnel; determining a product dripping speed based on the second image, and determining a speed difference based on a raw material dripping speed and the product dripping speed; generating an adjusting temperature according to a deviation value between the speed difference and a target speed range, generating a target temperature value according to the adjusting temperature, and adjusting the temperature of the heating plate according to the target temperature value; the reaction efficiency and the product yield can be improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology, and in particular to a method and system for intelligent control of organic synthesis reactions based on visual recognition. Background Art

[0002] In traditional organic synthesis experiments, liquid addition speed often relies on manual control or contact flow meters, which carries contamination risks and low precision. Reaction temperature regulation relies on manual experience, making it difficult to respond to changes in reaction conditions in real time. Existing technologies lack dynamic correlation analysis between raw material addition rate and product formation rate, limiting reaction efficiency and product yield. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for intelligent control of organic synthesis reactions based on visual recognition to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] In a first aspect, an embodiment of the present invention provides a method for intelligent control of organic synthesis reactions based on visual recognition, the method comprising the following steps:

[0006] Acquire a first image of the raw material dropping funnel and a second image of the product collecting port;

[0007] Identifying a change in the liquid level in the funnel based on the first image, and determining a dripping rate of the raw material based on the change in the liquid level in the funnel;

[0008] determining a product dripping rate based on the second image, and determining a speed difference based on the raw material dripping rate and the product dripping rate;

[0009] An adjustment temperature is generated according to a deviation value between the speed difference and the target speed range, a target temperature value is generated according to the adjustment temperature, and the temperature of the heating plate is adjusted according to the target temperature value.

[0010] Optionally, the calculation formula for adjusting the temperature is:

[0011] ΔT=Kp·ΔV+Ki·∫ΔVdt+Kd·dΔV / dt;

[0012] Among them, ΔT is the adjustment temperature, Kp is the proportional coefficient, Ki is the integral coefficient; Kd is the differential coefficient, and ΔV is the deviation value.

[0013] Optionally, the method further includes:

[0014] When the deviation value exceeds the deviation threshold for a continuous period of time that reaches the set time, the integral coefficient is increased;

[0015] When the fluctuation frequency of the deviation value is higher than the set frequency, the proportional coefficient is reduced and the differential coefficient is increased.

[0016] Optionally, the method further includes:

[0017] Acquire a solvent boiling point database, wherein the solvent boiling point database includes boiling point thresholds of a plurality of solvents;

[0018] When the real-time temperature of the heating plate reaches the boiling point threshold of the solvent in the funnel, the maximum output power of the heating plate is determined according to the target power of the heating plate, the real-time temperature and the boiling point of the solvent;

[0019] The output power of the heating plate is controlled to be lower than the maximum output power.

[0020] Optionally, identifying a liquid level change in the funnel based on the first image and determining a raw material dripping rate based on the liquid level change in the funnel includes:

[0021] Performing Gaussian filtering on the first image, performing Canny edge detection on the first image after Gaussian filtering, and outputting a droplet edge contour;

[0022] Extract the droplet edge contour based on HSV color space segmentation, perform morphological closing operation, and output the droplet area mask;

[0023] Perform connected domain analysis on the droplet region mask to calculate the pixel area of ​​the droplet region;

[0024] The pixel areas of the first images of multiple consecutive frames are integrated according to the droplet displacement to obtain the raw material dripping rate.

[0025] Optionally, the method further includes:

[0026] Extracting a funnel edge region in the first image based on HSV color space segmentation to generate a funnel edge image;

[0027] The circular outline of the funnel edge image is detected using Hough transform, the pixel diameter of the circular outline is determined, and the ratio of the funnel's calibration diameter to the pixel diameter is calculated.

[0028] determining a pixel offset of the funnel-shaped region based on the plurality of first images, and updating the ratio according to the ratio, the pixel offset, and a preset material thermal expansion coefficient;

[0029] A pixel-volume mapping model is established based on the updated ratio, the droplet volume is updated based on the pixel-volume mapping model and the pixel area, and the raw material dripping rate is updated based on the updated droplet volume.

[0030] Optionally, the updating formula of the ratio is: k'=k·(1+α·Δd); wherein k is the ratio of the nominal diameter of the funnel to the pixel diameter, α is a preset value of the thermal expansion coefficient of the material, and Δd is the pixel offset.

[0031] In a second aspect, an embodiment of the present invention provides an intelligent control system for organic synthesis reactions based on visual recognition, the system comprising:

[0032] at least one processor;

[0033] at least one memory for storing at least one program;

[0034] When the at least one program is executed by the at least one processor, the at least one processor implements any one of the methods described above.

[0035] The beneficial effects of the present invention are as follows: the present invention adopts non-contact visual droplet analysis: integrating dynamic cameras, edge detection algorithms and adaptive calibration, which is different from traditional contact or fixed parameter detection; adopts a multimodal intelligent control system: dynamically adjusts the temperature through speed differences to achieve closed-loop control of the reaction rate; the present invention can improve reaction efficiency and product yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention 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 of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0037] Figure 1 Schematic diagram of a flow chart of an intelligent control method for an organic synthesis reaction based on visual recognition in an embodiment of the present invention;

[0038] Figure 2 This is a control logic diagram for intelligent regulation of organic synthesis reactions in an embodiment of the present invention;

[0039] Figure 3 Schematic diagram of the structure of an intelligent control system for organic synthesis reactions based on visual recognition in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present invention so as to fully understand the purpose, scheme and effect of the present invention. It should be noted that the embodiments and features in the embodiments of the present invention can be combined with each other unless there is any conflict.

[0041] Among the related technologies, patents with publication numbers CN110665421A and CN110596418A focus on the detection of liquid flow rates in stirring equipment, using sensors and intelligent algorithms to optimize stirring parameters. However, their core technology is contact or built-in sensor technology, and the application scenario is industrial stirring equipment, which is significantly different from laboratory non-contact visual inspection. Patent publication number CN119000633A uses fluorescence to achieve non-contact detection of oil suspensions in water, and is applied in the field of water quality monitoring rather than laboratory synthesis. Ultrasonic Electronics' "Non-contact Pipe Scanning Equipment" patent (application number CN202422925935.2) performs non-destructive testing through liquid media, but its core is ultrasonic scanning of pipe defects, which has nothing to do with droplet volume analysis.

[0042] Existing utility model patents, such as the "Liquid Flow Rapid Detection Device" (application number CN202021919584.X), rely on sampling tubes and measuring containers, performing contact-based physical measurements and not involving visual dynamic analysis. While electronic micropipettes (such as BRAND's Transferpette) involve high-precision liquid manipulation, their functionality is limited to pipetting control and lacks integrated temperature control or visual analysis.

[0043] Existing organic synthesis experiments have the following pain points:

[0044] 1. The liquid dripping rate relies on manual stopcock adjustment, which has poor accuracy and cannot dynamically respond to the reaction state;

[0045] 2. The temperature control and the dripping process lack data linkage, which can easily lead to local overheating or incomplete reaction;

[0046] 3. Contact flow meters have contamination risks and are not suitable for miniaturized reaction devices;

[0047] 4. Multi-parameter coordination requires manual supervision, making it difficult to achieve long-term stable control;

[0048] The present invention makes the following improvements to the prior art:

[0049] 1. Non-contact visual recognition: Existing patents do not use dynamic cameras to analyze the funnel diameter and droplet volume. Traditional methods mostly use contact sensors or fixed parameter calculations.

[0050] 2. Multimodal collaborative control: Dual cameras monitor the flow rate difference (ΔV) between raw materials and products, and adjust the temperature through a Bluetooth-linked PID algorithm. This closed-loop control logic is not reflected in existing technologies.

[0051] 3. Adaptive compensation mechanism: The built-in solvent boiling point database and fuzzy PID dynamic parameter adjustment function enhance the intelligence and safety of the system, which is different from traditional fixed threshold control.

[0052] 4. Unique application scenarios: Existing patents mostly focus on industrial mixing, water quality monitoring, or pipe testing, while user patents target the small-scale, high-precision needs of laboratory organic synthesis, resolving the pain points of manual control contamination risks and low efficiency.

[0053] The present invention has applications in organic synthesis experimental technology, microfluidics, and pharmaceutical synthesis. In particular, the automated experimental device, based on visual recognition and Bluetooth communication, is used for contactless monitoring of liquid flow rates, dynamic adjustment of reaction temperatures, and coordinated control of raw material addition and product formation rates. It is suitable for synthetic reactions sensitive to addition velocity, and is particularly well-suited for remote control of hazardous experiments and high-throughput parallel synthesis experiments.

[0054] See Figure 1 The present invention provides a method for intelligent control of organic synthesis reactions based on visual recognition, the method comprising the following steps:

[0055] S100, acquiring a first image of a raw material dropping funnel and a second image of a product collecting port;

[0056] S200, identifying a change in the liquid level in the funnel based on the first image, and determining a dripping rate of the raw material based on the change in the liquid level in the funnel;

[0057] S300, determining a product dripping rate based on the second image, and determining a speed difference based on the raw material dripping rate and the product dripping rate;

[0058] S400 , generating an adjustment temperature according to a deviation value between the speed difference and the target speed range, generating a target temperature value according to the adjustment temperature, and adjusting the temperature of the heating plate according to the target temperature value.

[0059] Specifically, the control module compares the speed difference with the set speed threshold, dynamically calculates the target temperature and sends it to the heating plate to achieve closed-loop control.

[0060] refer to Figure 2 , the control logic is as follows:

[0061] Visual detection input: Camera 1: captures the first image of raw material dripping in real time and calculates the raw material dripping rate V1 (mL / min) through visual algorithm; Camera 2: monitors the product dripping rate V2 and transmits it to the control module synchronously.

[0062] The velocity difference (ΔV) is calculated using the formula ΔV = V1 - V2; the velocity difference ΔV is dynamically compared with the target velocity range set by the user (e.g., ΔV∈[0.5,1.2]mL / min).

[0063] In some embodiments, the calculation formula for adjusting the temperature is:

[0064] ΔT=Kp·ΔV+Ki·∫ΔVdt+Kd·dΔV / dt;

[0065] Among them, ΔT is the adjustment temperature, Kp is the proportional coefficient, Ki is the integral coefficient; Kd is the differential coefficient, and ΔV is the deviation value.

[0066] Specifically, the calculation formula of the PID controller is: ΔT = Kp·ΔV + Ki·∫ΔVdt + Kd·dΔV / dt; where ΔT is the regulated temperature, Kp is the proportional term used to quickly respond to the deviation value ΔV; Ki is the integral term used to eliminate steady-state errors; Kd is the differential term used to suppress oscillations, and ΔV is the deviation value.

[0067] Generate temperature instructions and communicate; generate target temperature value based on ΔT (such as T_new = T_current ± ΔT); send temperature instructions to the heating plate via BLE5.0 protocol.

[0068] Real-time temperature feedback: The heating plate’s built-in temperature sensor transmits data back for PID closed-loop correction and safety monitoring.

[0069] The drip rate-temperature transfer function model based on the dual cameras of the present invention is specifically:

[0070] Using non-contact visual monitoring of the flow rates of raw materials and products, the first camera continuously monitors the speed V1 of the raw material dripping in, and the second camera simultaneously monitors the speed V2 of the product dripping out. The control module calculates the real-time speed difference ΔV=V1-V2;

[0071] The embedded PID controller generates a temperature adjustment command ΔT based on the deviation between the speed difference ΔV and the target speed range set by the user, and dynamically adjusts the temperature of the heating plate via Bluetooth to achieve closed-loop control of ΔV and temperature.

[0072] Safety constraint module: Built-in solvent boiling point database. When the heating temperature approaches the current solvent boiling point, the temperature limit logic is automatically triggered, the heating power is preferentially reduced and an alarm signal is issued.

[0073] In some embodiments, the method further comprises:

[0074] When the deviation value exceeds the deviation threshold for a continuous period of time that reaches the set time, the integral coefficient is increased;

[0075] When the fluctuation frequency of the deviation value is higher than the set frequency, the proportional coefficient is reduced and the differential coefficient is increased.

[0076] Specifically, the parameter adjustment method of the PID controller includes:

[0077] When the speed difference ΔV exceeds the deviation threshold continuously for a set time period t1, the integral coefficient Ki is automatically increased to accelerate the elimination of the steady-state error.

[0078] When the speed difference ΔV fluctuation frequency is higher than the set value, the proportional coefficient Kp is automatically reduced and the differential coefficient Kd is increased to suppress temperature oscillation.

[0079] The dual-camera collaborative control method includes:

[0080] A timestamp synchronization mechanism is used to ensure that the calculation of the raw material drip rate V1 and the product drip rate V2 is based on the same time window;

[0081] If any camera data is lost, it automatically switches to single-camera mode and performs open-loop temperature control based only on the remaining camera data.

[0082] The present invention breaks through the existing technologies of single parameter control (such as only temperature or only dripping rate), open-loop control or single variable feedback (such as direct heating adjustment by temperature sensor); through dual variable collaborative feedback (speed difference ΔV as the core control quantity); and a double-layer safety logic combining dynamic PID parameter adjustment and solvent boiling point constraint.

[0083] In some embodiments, the method further comprises:

[0084] Acquire a solvent boiling point database, wherein the solvent boiling point database includes boiling point thresholds of a plurality of solvents;

[0085] When the real-time temperature of the heating plate reaches the boiling point threshold of the solvent in the funnel, the maximum output power of the heating plate is determined according to the target power of the heating plate, the real-time temperature and the boiling point of the solvent;

[0086] The output power of the heating plate is controlled to be lower than the maximum output power.

[0087] Temperature safety limit; built-in solvent boiling point database (such as ether: 34.6℃; toluene: 110.6℃); lower the boiling point by an appropriate temperature to obtain the boiling point threshold; when the real-time temperature gradually approaches the boiling point and reaches the boiling point threshold, the heating power is forced to be reduced and an alarm is triggered.

[0088] The safety constraint module further includes: according to the difference ΔT_safe between the boiling point of the solvent and the real-time temperature, the PID controller dynamically limits the maximum output power P_max of the heating plate, satisfying P_max=P_base·(1-ΔT_safe / T_boil), where P_base is the target power of the heating plate, ΔT_safe is the difference between the boiling point of the solvent and the real-time temperature, and T_boil is the boiling point of the solvent.

[0089] In some embodiments, in S200, identifying a change in the liquid level in the funnel based on the first image and determining a dripping rate of the raw material based on the change in the liquid level in the funnel includes:

[0090] S210, performing Gaussian filtering on the first image, performing Canny edge detection on the Gaussian filtered first image, and outputting a droplet edge contour;

[0091] S220, extracting the edge contour of the droplet based on HSV color space segmentation, performing a morphological closing operation, and outputting a droplet area mask;

[0092] S230, performing connected domain analysis on the droplet region mask to calculate the pixel area of ​​the droplet region.

[0093] S240 , integrating pixel areas of a plurality of consecutive first image frames according to the droplet displacement to obtain a raw material droplet velocity.

[0094] A specific embodiment is provided below to illustrate the method provided by the present invention:

[0095] During the esterification reaction, two cameras were installed to monitor the dripping rate of acetic anhydride and the dripping rate of ethyl acetate. When ΔV suddenly increased due to the exothermic reaction, the device automatically lowered the heating plate temperature from 110°C to 95°C, returning the reaction rate to the set range and increasing the yield by 18.7%. The steps are as follows:

[0096] First, magnetically attach the camera to the outer wall of the flask and complete initial calibration using the software (entering the nominal funnel diameter and selecting the solvent type). Next, set the target deviation range (e.g., 0.5-1.2 mL / min) and temperature control parameters (PID coefficients). Once enabled, the vision module captures images every 200 ms, extracts the droplet region using HSV color space segmentation, and calculates the instantaneous flow rate using a pixel-volume mapping model, as follows:

[0097] Input original image: The camera captures the droplet image (including background noise), and the camera installed on the outer wall of the flask captures the droplet falling process.

[0098] Apply a Gaussian filter to the original image. This filter reduces image noise and smoothes out interfering details through Gaussian kernel convolution. Parameters: σ = 1.5, kernel size 5×5.

[0099] Perform Canny edge detection on the filtered image and output the binary outline of the droplet edge, specifically:

[0100] Calculate image gradient (Sobel operator); non-maximum suppression; double threshold screening (low threshold = 50, high threshold = 150).

[0101] Droplet segmentation: Combine the Canny edge results with HSV color space threshold segmentation (for example, set H∈[20,100], S>40), and use morphological closing operations to fill internal holes. Output the droplet region mask.

[0102] Volume calculation:

[0103] A pixel-to-volume mapping model is established based on the funnel's calibrated diameter (user input) and the image pixel ratio;

[0104] Perform connected domain analysis on the droplet area mask and calculate the pixel area;

[0105] The droplet displacement between consecutive frames is integrated to generate the instantaneous flow rate of the raw material (mL / min), that is, the raw material dripping rate.

[0106] In some embodiments, the method further comprises:

[0107] Extracting a funnel edge region in the first image based on HSV color space segmentation to generate a funnel edge image;

[0108] The circular outline of the funnel edge image is detected using Hough transform, the pixel diameter of the circular outline is determined, and the ratio of the funnel's calibration diameter to the pixel diameter is calculated.

[0109] determining a pixel offset of the funnel-shaped region based on the plurality of first images, and updating the ratio according to the ratio, the pixel offset, and a preset material thermal expansion coefficient;

[0110] A pixel-volume mapping model is established based on the updated ratio, the droplet volume is updated based on the pixel-volume mapping model and the pixel area, and the raw material dripping rate is updated based on the updated droplet volume.

[0111] In some embodiments, the updating formula of the ratio is: k'=k·(1+α·Δd); wherein k is the ratio of the nominal diameter of the funnel to the pixel diameter, α is the preset value of the thermal expansion coefficient of the material, and Δd is the pixel offset.

[0112] Specifically, a non-invasive intelligent control device is used for image acquisition and temperature control, including:

[0113] 3D vision measurement module: A dual-spectral camera (visible light and near-infrared) is used to build a stereoscopic vision system. Multimodal image fusion technology is used to dynamically identify droplet morphological characteristics. Combined with a preset container parameter database, millimeter-level volume measurement accuracy is achieved.

[0114] Edge computing unit: The built-in NPU chip runs the lightweight YOLO-LITE model to analyze the droplet motion trajectory and calculate the instantaneous flow rate in real time;

[0115] Fuzzy PID control algorithm: establishes a dripping rate-temperature transfer function model and dynamically adjusts the heating power through the feedback speed difference;

[0116] Distributed communication architecture: uses BLE-Mesh protocol to achieve multi-device networking and supports 256-node synchronous control;

[0117] The device used in this embodiment includes the following modules:

[0118] Visual recognition module: A miniature high-definition camera (which may include a wide-angle lens, polarizing filter, and active ring-type fill-light LED array) with a built-in image processing chip dynamically captures the changes in the funnel liquid level and the product dripping image, and combines it with an edge detection algorithm to calculate the droplet volume and raw material dripping rate;

[0119] Control module: embedded microprocessor (such as ARMCortex-M7), integrating liquid flow rate calculation, temperature control logic and data communication functions;

[0120] Bluetooth communication module: supports BLE5.0 protocol and two-way communication with heating plate / heating sleeve;

[0121] Heating control module: generates temperature adjustment instructions through PID algorithm;

[0122] Display and interaction module: 1.8-inch LCD touch screen, supports parameter settings (monitoring interval, target speed difference threshold);

[0123] Power module: Rechargeable lithium battery, supports USB-C fast charging.

[0124] The software system includes:

[0125] Dynamic calibration protocol: Automatically identify the specifications of the flask through AR marking and establish the three-dimensional coordinate system of the container.

[0126] Adaptive learning algorithm: Builds a drip rate prediction model based on the LSTM network to continuously optimize control parameters.

[0127] Visualization interface: provides three-dimensional reaction process simulation view and multi-dimensional data correlation analysis.

[0128] Adaptive temperature compensation: Built-in database of common solvent boiling points. When the temperature is detected to be close to the boiling point of the solvent, the heating power is automatically reduced and an alarm is issued.

[0129] Dynamic calibration of funnel diameter: The user inputs the initial nominal diameter of the funnel through the interactive interface;

[0130] The visual recognition module is based on the funnel edge image captured in real time by the camera, and adopts an adaptive deformation compensation model (such as an edge distortion correction model based on a convolutional neural network) to dynamically correct the pixel-volume mapping model used for droplet volume calculation. The present invention breaks through the existing technology that relies on fixed calibration parameters or manual measurement, and is unable to adapt to the diameter changes caused by thermal deformation of the funnel during the experiment. Through the dual redundancy mechanism of the deformation compensation model and the user's initial calibration, "one-time calibration, continuous adaptation" is achieved.

[0131] and Figure 1 Corresponding to the method, refer to Figure 3 , an embodiment of the present invention provides an intelligent control system for organic synthesis reactions based on visual recognition, comprising:

[0132] at least one processor;

[0133] at least one memory for storing at least one program;

[0134] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0135] It can be seen that the contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0136] In addition, an embodiment of the present invention further discloses a computer program product or computer program, which is stored in a computer-readable storage medium. A processor of a computer device can read the computer program from the computer-readable storage medium, and the processor executes the computer program, causing the computer device to perform the above-mentioned method. Similarly, the contents of the above-mentioned method embodiment are applicable to the present storage medium embodiment. The functions specifically implemented by the present storage medium embodiment are the same as those of the above-mentioned method embodiment, and the beneficial effects achieved are also the same as those achieved by the above-mentioned method embodiment.

[0137] Those skilled in the art will appreciate that all or some of the methods and systems disclosed above can be implemented as software, firmware, hardware, or any suitable combination thereof. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As is well known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0138] The above is a specific description of the preferred implementation of the present disclosure, but the present disclosure is not limited to the above-mentioned implementation mode. Technical personnel familiar with the art can also make various equivalent modifications or substitutions without violating the spirit of the present disclosure. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present disclosure.

Claims

1. A method for intelligent control of organic synthesis reactions based on visual recognition, characterized in that: The method comprises the following steps: Acquire a first image of the raw material dropping funnel and a second image of the product collecting port; Identifying a change in the liquid level in the funnel based on the first image, and determining a dripping rate of the raw material based on the change in the liquid level in the funnel; determining a product dripping rate based on the second image, and determining a speed difference based on the raw material dripping rate and the product dripping rate; An adjustment temperature is generated according to a deviation value between the speed difference and the target speed range, a target temperature value is generated according to the adjustment temperature, and the temperature of the heating plate is adjusted according to the target temperature value.

2. The method according to claim 1, characterized in that The calculation formula for adjusting the temperature is: ΔT=Kp·ΔV+Ki·∫ΔVdt+Kd·dΔV / dt; Among them, ΔT is the adjustment temperature, Kp is the proportional coefficient, Ki is the integral coefficient; Kd is the differential coefficient, and ΔV is the deviation value.

3. The method according to claim 2, characterized in that The method further comprises: When the deviation value exceeds the deviation threshold for a continuous period of time that reaches the set time, the integral coefficient is increased; When the fluctuation frequency of the deviation value is higher than the set frequency, the proportional coefficient is reduced and the differential coefficient is increased.

4. The method according to claim 1, wherein The method further comprises: Acquire a solvent boiling point database, wherein the solvent boiling point database includes boiling point thresholds of a plurality of solvents; When the real-time temperature of the heating plate reaches the boiling point threshold of the solvent in the funnel, the maximum output power of the heating plate is determined according to the target power of the heating plate, the real-time temperature and the boiling point of the solvent; The output power of the heating plate is controlled to be lower than the maximum output power.

5. The method according to claim 1, wherein The identifying the change of the liquid level in the funnel based on the first image and determining the dripping rate of the raw material based on the change of the liquid level in the funnel includes: Performing Gaussian filtering on the first image, performing Canny edge detection on the first image after Gaussian filtering, and outputting a droplet edge contour; Extract the droplet edge contour based on HSV color space segmentation, perform morphological closing operation, and output the droplet area mask; Perform connected domain analysis on the droplet region mask to calculate the pixel area of ​​the droplet region; The pixel areas of the first images of multiple consecutive frames are integrated according to the droplet displacement to obtain the raw material dripping rate.

6. The method according to claim 5, characterized in that The method further comprises: Extracting a funnel edge region in the first image based on HSV color space segmentation to generate a funnel edge image; The circular outline of the funnel edge image is detected using Hough transform, the pixel diameter of the circular outline is determined, and the ratio of the funnel's calibration diameter to the pixel diameter is calculated. determining a pixel offset of the funnel-shaped region based on the plurality of first images, and updating the ratio according to the ratio, the pixel offset, and a preset material thermal expansion coefficient; A pixel-volume mapping model is established based on the updated ratio, the droplet volume is updated based on the pixel-volume mapping model and the pixel area, and the raw material dripping rate is updated based on the updated droplet volume.

7. The method according to claim 6, characterized in that The updating formula of the ratio is: k'=k·(1+α·Δd); wherein k is the ratio of the nominal diameter of the funnel to the pixel diameter, α is the preset value of the thermal expansion coefficient of the material, and Δd is the pixel offset.

8. An intelligent control system for organic synthesis reactions based on visual recognition, characterized in that: The system comprises: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

Images