Automatic liquid-liquid extraction method based on vision

Abstract

The invention provides an automatic liquid-liquid extraction method based on vision, and relates to the technical field of automatic extraction. The automatic liquid-liquid extraction method based on vision comprises the following steps: S1, three-dimensional modeling and liquid level function construction: carrying out three-dimensional modeling on a separating funnel, constructing a liquid level and volume function of the separating funnel, and determining the relationship between the liquid level height and the liquid volume through a mathematical model; s2, image acquisition and liquid level analysis; s3, liquid volume calculation; and S4, automatic control and liquid collection. By combining a traditional image processing method and a machine learning algorithm, the liquid level height and the boundary position of the upper and lower layers of liquid can be accurately recognized, the recognition precision is high, the robustness is high, and stable work can be achieved even under the condition of a complex liquid interface (for example, foam, impurities or refractive index changes exist).

Description

Technical Field

[0001] The present invention relates to the technical field of automatic extraction, and specifically to a vision-based automatic liquid-liquid extraction method. Background Art

[0002] Liquid-liquid extraction is a widely used separation technique, mainly used for separating and purifying target substances from mixtures. The traditional liquid-liquid extraction process usually relies on manual operation, and the valves of separating funnels are manually controlled to separate liquids with different densities. However, this method has many deficiencies:

[0003] 1. Manual operation error: When manually controlling the valves of separating funnels, it is difficult to precisely control the separation of liquids, which easily leads to incomplete liquid stratification or liquid mixing, affecting the extraction effect.

[0004] 2. Low efficiency: Manual operation takes a long time, especially when dealing with multiple samples, the efficiency is low, and it is difficult to meet the requirements of large-scale production or high-throughput experiments.

[0005] 3. Lack of real-time monitoring: The traditional method cannot monitor the liquid level change and the position of the demarcation line in real time, and it is difficult to achieve automatic and intelligent control.

[0006] 4. Poor repeatability: Due to the influence of human factors, the repeatability and stability of the traditional method are poor, and it is difficult to ensure the consistency of experimental results.

[0007] In recent years, with the development of automation technology and computer vision technology, automatic liquid-liquid extraction devices have gradually become a research hotspot. However, there is still a lack of an efficient liquid-liquid extraction method in the existing technology that can accurately identify the liquid level demarcation line, automatically calculate the liquid volume, and achieve automatic control. Summary of the Invention

[0008] (1) Technical Problems to be Solved

[0009] In view of the deficiencies of the existing technology, the present invention provides a vision-based automatic liquid-liquid extraction method, which solves the problems of low efficiency and low precision caused by the fact that the existing liquid-liquid extraction process usually relies on manual operation.

[0010] (2) Technical Solutions

[0011] To achieve the above objectives, the present invention is realized through the following technical solutions: A vision-based automatic liquid-liquid extraction method includes the following steps:

[0012] S1. Three-dimensional modeling and liquid level function construction: Perform three-dimensional modeling on the separating funnel, construct the liquid level and volume function of the separating funnel, and determine the relationship between the liquid level height and the liquid volume through a mathematical model;

[0013] S2. Image acquisition and liquid level analysis: Use a high-resolution camera to take pictures of the extraction liquid and extraction samples in the separatory funnel, and analyze the liquid level heights and liquid level dividing lines of the upper and lower layers in the image by combining traditional image processing methods (such as edge detection and binarization) with machine learning algorithms;

[0014] S3. Liquid volume calculation: According to the three-dimensional model of the separatory funnel and the liquid level function, and in combination with the distance between the upper layer liquid and the dividing line and the distance between the lower layer liquid and the separatory funnel valve, automatically calculate the volumes of the two layers of liquid;

[0015] S4. Automatic control and liquid collection: According to the calculated volume of the lower layer liquid, automatically control the rotation angle and opening and closing time of the electric valve of the separatory funnel to make the lower layer liquid flow out and be collected. Similarly, according to the volume of the upper layer liquid, control the valve to collect the upper layer liquid.

[0016] Preferably, in the step S1, three-dimensional modeling adopts laser scanning, structured light scanning or computer-aided design modeling technology, and the liquid level and volume function is obtained through experimental calibration or theoretical calculation.

[0017] Preferably, in the step S2, the high-resolution camera adopts an industrial-grade color or black-and-white camera, and the image acquisition frequency is dynamically adjusted according to the extraction process to ensure real-time capture of liquid level changes.

[0018] Preferably, in the step S2, the traditional image processing methods include grayscale conversion, filtering and denoising, edge detection, and contour extraction, and the machine learning algorithms include convolutional neural networks, support vector machines or deep learning models, which are used to improve the accuracy and robustness of liquid level dividing line recognition.

[0019] Preferably, in the step S3, the liquid volume calculation combines the liquid level function with the geometric parameters of the separatory funnel and adopts numerical integration or analytical solution methods, and the calculation accuracy reaches within 0.1%.

[0020] Preferably, in the step S4, the rotation angle and opening and closing time of the electric valve are precisely controlled by a PID controller or a fuzzy controller to adapt to the flow rate and viscosity characteristics of different liquids.

[0021] Preferably, it further includes step S5, real-time monitoring and feedback regulation: Real-time monitor the extraction process through the image acquisition system, and combine the liquid level changes to feedback and regulate the control parameters of the electric valve to ensure the accuracy and stability of liquid collection.

[0022] Preferably, the separatory funnel is made of transparent or semi-transparent material to improve the clarity of image acquisition.

[0023] (III) Beneficial effects

[0024] The present invention provides a vision-based automatic liquid-liquid extraction method, which has the following beneficial effects:

[0025] 1. By combining traditional image processing methods and machine learning algorithms, the present invention can accurately identify the liquid level heights of the upper and lower layers of liquids and the position of the liquid level boundary, with high recognition accuracy and strong robustness, and can work stably even in the case of complex liquid interfaces (such as the presence of foam, impurities, or refractive index changes).

[0026] 2. The present invention monitors the extraction process in real time through an image acquisition system, and adjusts the control parameters of the electric valve in combination with the feedback of the liquid level change to ensure the accuracy and stability of liquid collection. This real-time monitoring function not only improves the stability of the extraction process, but also can detect and correct abnormal situations in a timely manner.

[0027] 3. The automated design of the present invention can achieve multi-station synchronous operation, which is suitable for large-scale production or high-throughput experiments. At the same time, automated control reduces the interference of human factors and significantly improves the repeatability and stability of the extraction process. Specific embodiments

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0029] Embodiment:

[0030] The embodiment of the present invention provides a vision-based automatic liquid-liquid extraction method, which includes the following steps:

[0031] S1. Three-dimensional modeling and liquid level function construction: perform three-dimensional modeling on the separatory funnel, construct the liquid level and volume function of the separatory funnel, and determine the relationship between the liquid level height and the liquid volume through a mathematical model;

[0032] S2. Image acquisition and liquid surface analysis: use a high-resolution camera to take pictures of the extraction liquid and extraction samples in the separatory funnel, and analyze the liquid level heights of the upper and lower layers of liquids and the liquid level boundary in the image by combining traditional image processing methods (such as edge detection, binarization processing) and machine learning algorithms. The traditional image processing methods include grayscale conversion, filtering and denoising, edge detection, and contour extraction. The machine learning algorithms include convolutional neural networks, support vector machines, or deep learning models, which are used to improve the accuracy and robustness of liquid level boundary recognition;

[0033] S3. Liquid volume calculation: Based on the 3D model of the separating funnel and the liquid level function, combined with the distances from the upper liquid to the demarcation line and from the lower liquid to the separating funnel valve, automatically calculate the volumes of the two layers of liquid. The liquid volume calculation uses the liquid level function in combination with the geometric parameters of the separating funnel, and adopts numerical integration or analytical solution methods, with a calculation accuracy of within 0.1%.

[0034] S4. Automatic control and liquid collection: According to the calculated volume of the lower liquid, automatically control the rotation angle and opening / closing time of the electric valve of the separating funnel to make the lower liquid flow out and be collected. Similarly, according to the volume of the upper liquid, control the valve to collect the upper liquid. The rotation angle and opening / closing time of the electric valve are precisely controlled by a PID controller or a fuzzy controller to adapt to the flow rate and viscosity characteristics of different liquids.

[0035] S5. Real-time monitoring and feedback regulation: Real-time monitor the extraction process through an image acquisition system, and combine the liquid level changes to feedback and regulate the control parameters of the electric valve to ensure the accuracy and stability of liquid collection.

[0036] The separating funnel is made of transparent or semi-transparent material to improve the clarity of image acquisition.

[0037] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A vision-based automatic liquid-liquid extraction method, characterized in that: The following steps are involved: S1. Three-dimensional modeling and liquid level function construction: three-dimensional modeling of the separatory funnel is performed, the liquid level and volume function of the separatory funnel is constructed, and the relationship between the liquid level height and the liquid volume is determined through a mathematical model; S2, image acquisition and liquid level analysis, using a high-resolution camera to photograph the extract and the extracted sample in the separatory funnel, combining traditional image processing methods with machine learning algorithms to analyze the liquid level heights and liquid level boundaries of the upper and lower layers in the image; S3, liquid volume calculation, based on the three-dimensional model and liquid level function of the separatory funnel, combined with the distance between the upper liquid and the dividing line, and the distance between the lower liquid and the separatory funnel valve, automatically calculate the volume of the two layers of liquid; S4, automatic control and liquid collection, according to the calculated volume of the lower layer liquid, automatically control the rotation angle and opening and closing time of the electric valve of the separating funnel to make the lower layer liquid flow out and be collected. Similarly, according to the volume of the upper layer liquid, control the valve to collect the upper layer liquid.

2. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: In the step S1, three-dimensional modeling is performed using laser scanning, structured light scanning or computer-aided design modeling technology, and the liquid level and volume function is obtained through experimental calibration or theoretical calculation.

3. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: In step S2, the high-resolution camera adopts an industrial-grade color or black-and-white camera, and the image acquisition frequency is dynamically adjusted according to the extraction process.

4. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: In step S2, traditional image processing methods include graying, filtering and denoising, edge detection, and contour extraction, and machine learning algorithms include convolutional neural networks, support vector machines, or deep learning models.

5. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: In step S3, the liquid volume is calculated by combining the liquid level function with the geometric parameters of the separatory funnel, using numerical integration or analytical solution.

6. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: In step S4, the rotation angle and opening and closing time of the electric valve are accurately controlled by a PID controller or a fuzzy controller.

7. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: The method also includes real-time monitoring and feedback adjustment in step S5, in which the extraction process is monitored in real time by an image acquisition system, and the control parameters of the electric valve are adjusted in combination with the feedback of the liquid level change.

8. The vision-based automatic liquid-liquid extraction method according to claim 1, characterized in that: The separating funnel is made of transparent or translucent material.