Automatic demulsification control algorithm based on emulsification state recognition

Abstract

The invention discloses an automatic demulsification control algorithm based on emulsification state recognition. The automatic demulsification control algorithm comprises the following steps: collecting video stream data after a liquid chemical reaction; extracting an ROI region through a deep learning segmentation network; classifying the ROI through a deep learning classification network to obtain a liquid emulsification judgment result; segmenting the liquid emulsification judgment result through a deep learning segmentation algorithm to obtain a specific emulsification layer area; calculating the proportion of the emulsion layer area image in the ROI area; calculating demulsifier parameters according to the proportion; calculating the stirring speed according to the proportion; and carrying out demulsification operation according to the demulsifier parameters and the stirring speed. The content of the demulsifier and the stirring speed are controlled in real time through the deep learning segmentation network and the deep learning classification network so as to adapt to different experiment scenes. And through self-feedback setting, high efficiency and reliability of the system are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to an automatic demulsification control algorithm based on emulsification state recognition. Background Art

[0002] In chemical experiments, when two immiscible liquids coexist, the presence of surfactants can cause an emulsion to form, blurring the previously clear interface. This phenomenon is known as demulsification. To restore the stratified state, demulsification is performed by adding a demulsifier or extending the standing time.

[0003] Demulsification is crucial in chemical experiments, directly impacting the smooth progress of subsequent key operations such as liquid separation and transfer. Failure to effectively break the emulsion can severely disrupt the experimental process, hindering the accurate separation and extraction of target components, and ultimately affecting the reliability and validity of experimental results.

[0004] However, in the field of demulsification in chemical experiments, traditional demulsification control methods currently rely primarily on subjective manual judgment. This highly subjective approach makes it difficult to precisely control the demulsification process. The varying experience levels of different operators can easily lead to significant discrepancies in demulsification results. Furthermore, manual judgment cannot track the dynamic changes in the emulsification state in real time. Lacking precise data support, it is difficult to flexibly adjust key parameters such as demulsifier dosage and stirring speed based on actual conditions. This results in low demulsification efficiency and makes it difficult to guarantee the accuracy of the final experimental results. This significantly restricts the high-quality conduct of chemical experiments and the reliability of research results. Summary of the Invention

[0005] The present invention provides an automatic demulsification control algorithm based on emulsification state recognition to address the problem that traditional demulsification control methods in the prior art mainly rely on manual subjective judgment. This method is highly subjective and difficult to achieve precise control of the demulsification process. Different operators have different levels of experience, which can easily lead to significant differences in demulsification results. Manual judgment cannot track the dynamic changes of the emulsification state in real time. The lack of accurate data as support makes it difficult to flexibly adjust core parameters such as demulsifier dosage and stirring speed according to actual conditions, resulting in low demulsification efficiency and difficulty in ensuring the accuracy of the final experimental results.

[0006] On the one hand, an embodiment of the present invention provides an automatic demulsification control algorithm based on emulsification state recognition, comprising:

[0007] Collect video stream data after liquid chemical reaction;

[0008] Extract ROI area through deep learning segmentation network;

[0009] Classifying the ROI area through a deep learning classification network to obtain a liquid emulsification determination result;

[0010] Segmenting the liquid emulsification determination result by a deep learning segmentation algorithm to obtain a specific emulsion layer area;

[0011] Calculating the proportion of the emulsion layer area image in the ROI area;

[0012] Calculating demulsifier parameters according to the proportions;

[0013] Calculate the stirring speed according to the proportion;

[0014] The demulsification operation is performed according to the demulsifier parameters and the stirring speed.

[0015] In one possible implementation, extracting the ROI region by using a deep learning segmentation network includes:

[0016] The video frame data of the video stream data is segmented and the background is removed by using the YOLOv8-Seg deep learning segmentation network to obtain the ROI region of interest.

[0017] In a possible implementation, classifying the ROI region by a deep learning classification network to obtain a liquid emulsification determination result includes:

[0018] The ROI area is classified using the YOLOv8-cls deep learning classification network to obtain an emulsified state result and a non-emulsified state result;

[0019] The demulsification completion result is obtained according to the non-emulsified state result.

[0020] In a possible implementation, segmenting the liquid emulsification determination result by a deep learning segmentation algorithm to obtain a specific emulsion layer area includes:

[0021] The SegFormer semantic segmentation network is used to perform regional segmentation on the ROI area of ​​the emulsified state result to obtain an emulsified layer area image.

[0022] In a possible implementation, calculating the proportion of the emulsion layer area image in the ROI area includes:

[0023] The proportion is calculated according to the pixel area of ​​the emulsion layer area image in the ROI area.

[0024] In a possible implementation, calculating demulsifier parameters according to the proportion includes:

[0025] The emulsifier addition amount of the demulsifier is calculated according to the area ratio of the emulsion layer area and the ROI area.

[0026] In a possible implementation, calculating the stirring speed according to the proportion includes:

[0027] The stirring speed required for demulsification is calculated according to the area ratio of the emulsion layer area and the ROI area.

[0028] In a possible implementation, the demulsification operation according to the demulsifier parameters and the stirring speed includes:

[0029] allowing the liquid to stand;

[0030] Return to the step of collecting the video stream data and perform detection again.

[0031] The automatic demulsification control algorithm based on emulsification state recognition in the present invention has the following advantages:

[0032] (1) The demulsifier content and stirring speed are controlled in real time through deep learning segmentation network and deep learning classification network to adapt to different experimental scenarios.

[0033] (2) The self-feedback setting ensures the efficiency and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] 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 or the description of the prior art. 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.

[0035] Figure 1 This is a flowchart of an automatic demulsification control algorithm based on emulsification state recognition provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] Figure 1A schematic flow chart of an automatic demulsification control algorithm based on emulsification state identification provided by an embodiment of the present invention; an automatic demulsification control algorithm based on emulsification state identification provided by an embodiment of the present invention includes:

[0038] Collect video stream data after liquid chemical reaction;

[0039] Extract ROI area through deep learning segmentation network;

[0040] Classifying the ROI area through a deep learning classification network to obtain a liquid emulsification determination result;

[0041] Segmenting the liquid emulsification determination result by a deep learning segmentation algorithm to obtain a specific emulsion layer area;

[0042] Calculating the proportion of the emulsion layer area image in the ROI area;

[0043] Calculating demulsifier parameters according to the proportions;

[0044] Calculate the stirring speed according to the proportion;

[0045] The demulsification operation is performed according to the demulsifier parameters and the stirring speed.

[0046] Extracting the ROI region by deep learning segmentation network includes:

[0047] The video frame data of the video stream data is segmented and the background is removed by using the YOLOv8-Seg deep learning segmentation network to obtain the ROI region of interest.

[0048] The ROI area is classified by a deep learning classification network to obtain the liquid emulsification determination results including:

[0049] The ROI area is classified using the YOLOv8-cls deep learning classification network to obtain an emulsified state result and a non-emulsified state result;

[0050] The demulsification completion result is obtained according to the non-emulsified state result.

[0051] The specific emulsion layer area obtained by segmenting the liquid emulsification determination result using a deep learning segmentation algorithm includes:

[0052] The SegFormer semantic segmentation network is used to perform regional segmentation on the ROI area of ​​the emulsified state result to obtain an emulsified layer area image.

[0053] Calculating the proportion of the emulsion layer area image in the ROI area includes:

[0054] The proportion is calculated according to the pixel area of ​​the emulsion layer area image in the ROI area.

[0055] Calculating the demulsifier parameters according to the proportions includes:

[0056] The emulsifier addition amount of the demulsifier is calculated according to the area ratio of the emulsion layer area and the ROI area.

[0057] Calculating the stirring speed according to the proportion includes:

[0058] The stirring speed required for demulsification is calculated according to the area ratio of the emulsion layer area and the ROI area.

[0059] After the demulsification operation is performed according to the demulsifier parameters and the stirring speed, the following steps are included:

[0060] allowing the liquid to stand;

[0061] Return to the step of collecting the video stream data and perform detection again.

[0062] For example, a fixed industrial camera is used to collect video stream data of liquid-liquid chemical reactions in real time.

[0063] The YOLOv8-Seg semantic segmentation network is used to extract regions of interest (ROIs) from video frames, effectively removing irrelevant background information and improving detection accuracy and efficiency.

[0064] The YOLOv8-Seg is a model in the YOLOv8 series that integrates target detection and instance segmentation functions. Its structure includes Backbone (C2f module and SPPF module extract multi-scale features), Neck (PAN-FPN structure fuses multi-scale features) and Head (detection head + segmentation head).

[0065] The YOLOv8-cls deep learning classification network is used to classify the extracted ROI area to determine whether it is in an emulsified state or a non-emulsified state.

[0066] YOLOv8-cls is a lightweight model in the YOLOv8 series designed specifically for image classification. Its architecture consists of Backbone (feature extraction), Neck (multi-scale fusion, which can be simplified or omitted in classification tasks), and Head (classification output). The Backbone uses the C2f module instead of the traditional C3 module, optimizing gradient flow to reduce redundant parameters and combining depthwise separable convolution to improve efficiency. The Head uses global average pooling (GAP) and fully connected layers to output class probabilities and supports the anchor-free mechanism, eliminating the need for pre-set anchor boxes.

[0067] If the YOLOv8-cls deep learning classification network identifies the solution as non-emulsified, the demulsification is determined to be complete and no emulsion layer exists in the solution; if the solution is identified as emulsified, the next step is entered.

[0068] For the ROI area identified as emulsified, the SegFormer semantic segmentation network is used to further segment the specific area of ​​the emulsified layer, providing a basis for subsequent quantitative analysis.

[0069] SegFormer is a semantic segmentation model based on the Transformer architecture. It consists of an encoder (MixTransformer, which extracts multi-scale features in layers) and a decoder (All-MLP, which fuses features only through linear layers). Its core design includes Overlap Patch Embedding (which preserves local context) and Efficient Self-Attention. With its extremely low parameter count, it is suitable for high-precision segmentation in complex scenes.

[0070] Among them, the model configuration of the SegFormer model is shown in the following code:

[0071]

[0072] Calculate the ratio of the pixel area of ​​the emulsion layer to the liquid pixel area of ​​the entire ROI area. The formula is as follows:

[0073]

[0074] Where, is the segmented emulsion layer area. W and H are the width and height of the ROI area, respectively. R is the area ratio.

[0075] The dosage of the demulsifier is dynamically adjusted by the area ratio R. The formula is as follows:

[0076]

[0077] Where, is the amount of emulsifier added. t is the frame position of the current video frame. represents the preset initial demulsifier content. represents the area ratio calculated at the tth frame.

[0078] The stirring speed is dynamically adjusted by the area ratio R. The formula is as follows:

[0079]

[0080] Where, is the stirring speed. t is the frame position of the current video frame. represents the preset initial stirring speed. represents the area ratio calculated at the tth frame.

[0081] Based on the demulsifier dosage and stirring speed obtained above, control the stirring equipment. After standing for N seconds, return to the initial step and continue monitoring the liquid state until demulsification is complete.

[0082] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including all changes and modifications that fall within the scope of the present invention and the preferred embodiments.

[0083] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An automatic demulsification control algorithm based on emulsification state recognition, characterized in that: include: Collect video stream data after liquid chemical reaction; Extract ROI area through deep learning segmentation network; Classifying the ROI area through a deep learning classification network to obtain a liquid emulsification determination result; Segmenting the liquid emulsification determination result by a deep learning segmentation algorithm to obtain a specific emulsion layer area; Calculating the proportion of the emulsion layer area image in the ROI area; Calculating demulsifier parameters according to the proportions; Calculate the stirring speed according to the proportion; The demulsification operation is performed according to the demulsifier parameters and the stirring speed.

2. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: Extracting the ROI region by deep learning segmentation network includes: The video frame data of the video stream data is segmented and the background is removed by using the YOLOv8-Seg deep learning segmentation network to obtain the ROI region of interest.

3. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: The ROI area is classified by a deep learning classification network to obtain the liquid emulsification determination results including: The ROI area is classified using the YOLOv8-cls deep learning classification network to obtain an emulsified state result and a non-emulsified state result; The demulsification completion result is obtained according to the non-emulsified state result.

4. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: The specific emulsion layer area obtained by segmenting the liquid emulsification determination result using a deep learning segmentation algorithm includes: The SegFormer semantic segmentation network is used to perform regional segmentation on the ROI area of ​​the emulsified state result to obtain an emulsified layer area image.

5. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: Calculating the proportion of the emulsion layer area image in the ROI area includes: The proportion is calculated according to the pixel area of ​​the emulsion layer area image in the ROI area.

6. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: Calculating the demulsifier parameters according to the proportions includes: The emulsifier addition amount of the demulsifier is calculated according to the area ratio of the emulsion layer area and the ROI area.

7. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: Calculating the stirring speed according to the proportion includes: The stirring speed required for demulsification is calculated according to the area ratio of the emulsion layer area and the ROI area.

8. The automatic demulsification control algorithm based on emulsification state recognition according to claim 1 is characterized in that: After the demulsification operation is performed according to the demulsifier parameters and the stirring speed, the following steps are included: allowing the liquid to stand; Return to the step of collecting the video stream data and perform detection again.

Images