Method for intelligently judging standing layering state of chemical experiment

Abstract

The invention provides a method for intelligently judging a standing layering state of a chemical experiment, which is applied to the technical field of chemical experiments and comprises the following steps: collecting experiment video stream data in real time; using a Yolov8 target detection algorithm to accurately identify a liquid layering area, and judging whether a layering phenomenon occurs or not; if the layering phenomenon occurs, judging whether an emulsion layer exists or not by adopting a Yolo-cls image classification technology; if the emulsion layer does not exist, accurately extracting an ROI (Region of Interest) of liquid level layering according to a Yolov8 target detection result; utilizing a DINOv2 model to record the depth semantic features of the liquid level layering area at different moments in real time; acquiring a similarity difference value at different moments by using a similarity algorithm; judging whether the difference value is within a preset threshold range or not; if yes, layering of the liquid level is completed; the prior information of the liquid level position needs to be reflected and known through equipment, and the liquid level can only be placed at a fixed position every time and cannot be judged in real time; in the prior art, the phenomenon that an emulsion layer appears in liquid is not considered, and real-time detection in the whole experiment process cannot be realized based on the liquid standing completion state.

Description

Technical Field

[0001] The present application relates to the technical field of chemical experiments, and in particular to a method for intelligently judging the static stratification state of a chemical experiment. Background Art

[0002] With the continuous development of intelligent technology, the traditional manufacturing industry is undergoing a profound renewal process. This change covers the transition from manual operation to automated production, and further moves towards a new stage of intelligent manufacturing. In traditional fields such as fine chemicals and pharmaceuticals, many experimental links have the possibility of replacing direct operations of scientific researchers with machine vision technology. In the automated experimental device of the experimental platform, there is a process where the liquid needs to be observed for static stratification, which is of great significance for subsequent judgment of the reaction progress, optimization of product separation and purification steps, in-depth exploration of the reaction mechanism, and ensuring experimental safety. In a closed experimental device, how to accurately monitor whether the liquid in the reaction bottle has completed stratification and how to strike a balance between real-time performance and accuracy is a difficult challenge.

[0003] The currently known patent CN117269071A proposes a method for judging the state of liquid surface emulsified fuzzy stratified extraction. The method uses a camera to obtain image data of the liquid surface emulsified layer at different times, and then uses an algorithm to obtain the corresponding color difference time curve. When the curve changes relatively gently at the liquid surface position and reaches the critical value set by the system, it is determined that emulsification occurs. However, this method requires prior information on the liquid surface position.

[0004] Patent CN118411520A proposes an intelligent recognition method for liquid level stratification in a columnar container. The method uses multiple cameras to obtain image data at different viewing angles, and then uses the spectrum to identify the position of liquid level stratification within the entire height range of the columnar container. The method does not take into account the phenomenon of emulsified layers in the liquid, and the method is based on the liquid being left to stand still, and cannot achieve real-time detection of the entire experimental process.

[0005] Patent CN116147731A proposes a liquid level detection method for a stratified liquid surface interface. This method obtains pressure information under different liquids through a TIP probe head. When the density of the second layer of liquid changes, the TIP probe will jump, thereby detecting the stratified liquid surface of the liquid. This method does not take into account the emulsification phenomenon.

[0006] To sum up, there are two obvious disadvantages:

[0007] 1. The prior information of the liquid level position needs to be reflected by the equipment. It can only be placed in a fixed position each time and cannot be judged in real time;

[0008] 2. The phenomenon of emulsified layer in the liquid is not taken into consideration, and the experiment is carried out based on the liquid being in a static state, so real-time detection of the entire experimental process cannot be achieved. Summary of the invention

[0009] The present application aims to solve the technical problems that the prior information of the position of the liquid surface needs to be reflected by the equipment, and it can only be placed in a fixed position each time, and it cannot be judged in real time. In addition, it does not take into account the phenomenon of emulsion layer of the liquid, and it is based on the liquid being in a static state, and it is impossible to realize real-time detection of the whole experiment process. A method for intelligently judging the static stratification state of a chemical experiment is provided.

[0010] This application adopts the following technical means to solve the technical problem: a method for intelligently judging the static stratification state of a chemical experiment,

[0011] A method for intelligently judging the static stratification state of a chemical experiment, the method comprising:

[0012] Collect experimental video stream data in real time;

[0013] Use Yolov8 target detection algorithm to accurately identify liquid stratification areas and determine whether stratification occurs;

[0014] If stratification occurs, Yolo-cls image classification technology is used to determine whether there is an emulsified layer;

[0015] If there is no emulsion layer, the ROI area of ​​the liquid surface stratification is accurately extracted according to the Yolov8 target detection results;

[0016] The DINOv2 model is used to record the deep semantic features of the liquid surface stratification areas at different times in real time;

[0017] Compare the three similarities of the features at the previous specific moment with the features at the current moment, and determine the difference between the maximum and minimum values ​​of the three similarities;

[0018] Determine whether the difference is within a preset threshold range;

[0019] If so, the liquid level has completed stratification.

[0020] Furthermore, in the step of using the Yolov8 target detection algorithm to accurately identify the liquid stratification area and determine whether stratification occurs,

[0021] If no delamination occurs, continue to perform real-time detection until delamination occurs.

[0022] Furthermore, in the step of using Yolo-cls image classification technology to determine whether an emulsified layer exists if stratification occurs,

[0023] If an emulsion layer exists, add a demulsifier or extend the standing time until the emulsion layer disappears completely.

[0024] Furthermore, before the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment to determine the difference between the maximum and minimum values ​​of the three similarities;

[0025] The depth semantic features of the liquid surface stratification area at the current moment are compared with those of the previous specific moment 10 seconds, 20 seconds or 30 seconds ago.

[0026] Furthermore, in the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment and determining the difference between the maximum and minimum values ​​of the three similarities;

[0027] The three similarities are Euclidean distance, Manhattan distance and cosine similarity.

[0028] Further, in the step of determining whether the difference is within a preset threshold range;

[0029] According to the Yolov8 target detection results, the ROI area of ​​the liquid surface stratification is accurately extracted.

[0030] Furthermore, before the step of collecting experimental video stream data in real time;

[0031] The mask processing technology is used to retain the reaction bottle area and eliminate unnecessary background interference.

[0032] The present application provides a method for intelligently judging the static stratification state of a chemical experiment, which has the following beneficial effects:

[0033] The Yolov8 algorithm is used to realize real-time detection of the liquid stratification interface in the reaction bottle and automatic judgment of the liquid emulsion layer. The three similarity algorithms are used to realize the detection of liquid static stratification. The overall solution detection results are more efficient, accurate, and real-time.

[0034] As a pure visual solution, it is completely based on the image data captured by the camera to achieve real-time liquid static monitoring and experimental process analysis, completely abandoning the dependence on traditional sensors and other hardware facilities, and realizing a remote, non-contact, unmanned and automated liquid static stratification technology, thereby reducing the harm of chemicals to scientific researchers. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flowchart of an embodiment of a method for intelligently judging the static stratification state of a chemical experiment of the present application;

[0036] Figure 2A schematic diagram of an RSTP captured video frame image of an embodiment of the method for intelligently judging the static stratification state of a chemical experiment of the present application;

[0037] Figure 3 This is a schematic diagram of an image of a reaction product region after RSTP cropping in one embodiment of the method for intelligently judging the static stratification state of a chemical experiment of the present application;

[0038] Figure 4 A schematic diagram of Yolov8 stratified liquid level detection results of one embodiment of the method for intelligently judging the static stratification state of a chemical experiment of the present application;

[0039] Figure 5 This is a schematic diagram of ROI region extraction in one embodiment of the method for intelligently judging the static stratification state of a chemical experiment in the present application.

[0040] The implementation, functional features and advantages of the present application will be further described in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0041] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0042] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0043] It should be noted that the terms "include", "comprises" and "have" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices. In the claims, specification and drawings of the present application, relational terms such as "first" and "second" are merely used to distinguish one entity / operation / object from another entity / operation / object, and do not necessarily require or imply any such actual relationship or order between these entities / operations / objects.

[0044] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0045] Reference Figure 1-5 , which is a flow chart of a method for intelligently judging the static stratification state of a chemical experiment in one embodiment of the present application;

[0046] Embodiment 1

[0047] A method for intelligently judging the static stratification state of a chemical experiment, the method comprising:

[0048] Collect experimental video stream data in real time;

[0049] Use Yolov8 target detection algorithm to accurately identify liquid stratification areas and determine whether stratification occurs;

[0050] If stratification occurs, Yolo-cls image classification technology is used to determine whether there is an emulsified layer;

[0051] If there is no emulsion layer, the ROI area of ​​the liquid surface stratification is accurately extracted according to the Yolov8 target detection results;

[0052] The DINOv2 model is used to record the deep semantic features of the liquid surface stratification areas at different times in real time;

[0053] Compare the three similarities of the features at the previous specific moment with the features at the current moment, and determine the difference between the maximum and minimum values ​​of the three similarities;

[0054] Determine whether the difference is within a preset threshold range;

[0055] If so, the liquid level has completed stratification.

[0056] In the step of using the Yolov8 target detection algorithm to accurately identify the liquid stratification area and determine whether stratification occurs,

[0057] If no delamination occurs, continue to perform real-time detection until delamination occurs.

[0058] In this embodiment, before the step of acquiring experimental video stream data in real time;

[0059] The mask processing technology is used to retain the reaction bottle area and eliminate unnecessary background interference;

[0060] Specifically, RSTP (Real-Time Streaming Protocol) technology is used to collect video data in the reaction bottle from the camera in real time. The data mainly describes the static process of the liquid in the reaction bottle, captures key phenomena such as liquid surface stratification, emulsion layer formation and disappearance, and provides high-quality input data for subsequent target detection, image classification and deep semantic feature extraction;

[0061] According to Figure 2 and Figure 3 Judging, removing the background from the video frame image with the background;

[0062] Subsequently, the Yolov8 target detection algorithm is used to detect the liquid stratification interface in the target area. If Yolov8 can detect the liquid stratification area, it means that the liquid has been initially stratified; otherwise, if the stratified liquid surface cannot be detected, it means that the liquid is still in a turbid state and no obvious stratification has been formed.

[0063] The target detection algorithm results are as follows Figure 4 .

[0064] In this embodiment, if stratification occurs, the Yolo-cls image classification technology is used to determine whether there is an emulsified layer.

[0065] If an emulsion layer exists, add a demulsifier or extend the standing time until the emulsion layer disappears completely.

[0066] Specifically, after the preliminary confirmation of liquid stratification, the liquid in the main reaction bottle is classified by Yolo-cls image classification technology to determine whether there is an emulsified layer;

[0067] If the emulsion layer disappears, it is necessary to determine the ROI area of ​​the liquid surface stratification, ROI is Region of Interest, region of interest;

[0068] The usual technique is to position the device at a certain position to capture image information. However, the position and color of the liquid surface separation in different tests are different. Therefore, a positioning action is required to determine the image area of ​​the liquid surface stratification, that is, the ROI area, and perform segmented analysis and determination.

[0069] According to Yolo v 8 The result of target detection determines the center point of the detection frame. Then, based on the given threshold, the ROI area of ​​the liquid surface layer is determined. The formula is as follows:

[0070] box roi =(x min ,y min , x max ,ymax )=(x c -T roi ,y c -T roi , x c +T roi ,y c +T roi )

[0071] Among them, you can refer to Figure 5 The ROI area is shown in

[0072] The DINOv2 model is used to obtain the deep semantic features of the liquid stratification area at different times. The formula is as follows:

[0073] f t =M dinv2 I roi )t=t 10s , t 20s , t 30s , t0

[0074] In the formula, M dinv2 () represents the DINOv2 model, I roi represents the ROI area of ​​the liquid layer, and t represents different moments, where t0 is the current moment and t 10s , t 20s , t 30s The first 10 seconds before t0, the first 20 seconds before t0 and the first 30 seconds before t0;

[0075] In this embodiment, before the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment and determining the difference between the maximum value and the minimum value of the three similarities;

[0076] The depth semantic features of the liquid surface stratification area at the current moment are compared with those of the previous specific moment 10 seconds, 20 seconds or 30 seconds ago.

[0077] In the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment, and determining the difference between the maximum and minimum values ​​of the three similarities;

[0078] The three similarities are Euclidean distance, Manhattan distance and cosine similarity.

[0079] In the step of determining whether the difference is within a preset threshold range;

[0080] According to the Yolov8 target detection results, the ROI area of ​​the liquid surface stratification is accurately extracted.

[0081] Based on the acquired deep semantic feature set, three different similarities (Euclidean distance, Manhattan distance and cosine similarity) between the current moment and all previous specific moment features are calculated, and the difference between the maximum and minimum values ​​of these similarity values ​​is determined;

[0082] The formula is as follows:

[0083]

[0084]

[0085] In the formula S is a similarity algorithm, including Euclidean distance S edu , Manhattan distance S man And cosine similarity S cos . R edu max , R man max and R cos max are the maximum values ​​of the Euclidean distance, Manhattan distance, and cosine similarity between the current moment and all previous moments. edu min , R man min and R cos min are the minimum values ​​of the Euclidean distance, Manhattan distance, and cosine similarity between the current moment and all previous moments. edu_diff , R man_diff and R cos_diff are the differences between the maximum and minimum values ​​of Euclidean distance, Manhattan distance, and cosine similarity, indicating the degree of change in similarity;

[0086] Next, set the Euclidean distance threshold T edu , Manhattan distance threshold T man and cosine similarity threshold T cos If the similarity difference exceeds the preset threshold T edu , T man and T cos , it indicates that the change is still drastic and the static stratification has not been completed; on the contrary, if the difference is within the threshold range, it indicates that the change is small and the static stratification is completed. In this scheme, the maximum voting method is used for the fusion of the three similarity algorithms. If at least two of the three similarity algorithms are confirmed to be in the static stratification state, it means that the static stratification process has been completed. Otherwise, it indicates that the current static stratification has not yet reached the completion state. The specific formula is as follows:

[0087]

[0088] Where Major vote() indicates the maximum voting method. Status 1 indicates that the liquid static stratification is completed. Status 0 indicates that the liquid static stratification is not yet completed. When the status is displayed as unfinished, it is necessary to return to the target detection ROI area and continue to perform feature extraction and similarity calculation processes until the static stratification is successfully completed;

[0089] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the process in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0091] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0092] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0093] Although the embodiments of the present application have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for intelligently judging the static stratification state of a chemical experiment, characterized in that: The method comprises: Collect experimental video stream data in real time; Use Yolov8 target detection algorithm to accurately identify liquid stratification areas and determine whether stratification occurs; If stratification occurs, Yolo-cls image classification technology is used to determine whether there is an emulsified layer; If there is no emulsion layer, the ROI area of ​​the liquid surface stratification is accurately extracted according to the Yolov8 target detection results; The DINOv2 model is used to record the deep semantic features of the liquid surface stratification areas at different times in real time; Compare the three similarities of the features at the previous specific moment with the features at the current moment, and determine the difference between the maximum and minimum values ​​of the three similarities; Determine whether the difference is within a preset threshold range; If so, the liquid level has completed stratification.

2. The method for intelligently judging the static stratification state of a chemical experiment according to claim 1, characterized in that: In the step of using the Yolov8 target detection algorithm to accurately identify the liquid stratification area and determine whether stratification occurs, If no delamination occurs, continue to perform real-time detection until delamination occurs.

3. The method for intelligently judging the static stratification state of a chemical experiment according to claim 1, characterized in that: In the step of using Yolo-cls image classification technology to determine whether there is an emulsified layer if stratification occurs, If an emulsion layer exists, add a demulsifier or extend the standing time until the emulsion layer disappears completely.

4. The method for intelligently judging the static stratification state of a chemical experiment according to claim 1, characterized in that: Before the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment to determine the difference between the maximum and minimum values ​​of the three similarities; The depth semantic features of the liquid surface stratification area at the current moment are compared with those of the previous specific moment 10 seconds, 20 seconds or 30 seconds ago.

5. The method for intelligently judging the static stratification state of a chemical experiment according to claim 4, characterized in that: In the step of comparing the three similarities of the features at the previous specific moment with the features at the current moment, and determining the difference between the maximum and minimum values ​​of the three similarities; The three similarities are Euclidean distance, Manhattan distance and cosine similarity.

6. The method for intelligently judging the static stratification state of a chemical experiment according to claim 5, characterized in that: In the step of determining whether the difference is within a preset threshold range; According to the Yolov8 target detection results, the ROI area of ​​the liquid surface stratification is accurately extracted.

7. The method for intelligently judging the static stratification state of a chemical experiment according to claim 1, characterized in that: Before the step of collecting experimental video stream data in real time; The mask processing technology is used to retain the reaction bottle area and eliminate unnecessary background interference.

Images