Fermentation defoaming control system, method and device based on machine vision

Abstract

The invention provides a fermentation defoaming control system, method and device based on machine vision, the system comprises: a fermentation tank, the fermentation tank comprises a tank body and a seal head installed at the top of the tank body, and the seal head is provided with a sight glass; the image acquisition device is mounted on the sight glass and is used for acquiring a current image in the tank body through the sight glass; the defoaming agent source is communicated with the fermentation tank through a pipeline, and a switch valve is mounted on the pipeline; the controller is used for receiving the current image acquired by the image acquisition device and comparing the current image with a pre-stored reference image to generate a comparison result; the controller is further used for generating an opening instruction under the condition that the comparison result is that the foam is excessive, and the opening instruction is used for controlling the switching valve to be opened. The technical problems that in the prior art, a defoaming process is poor in monitoring effect and high in misjudgment rate are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent equipment, and in particular to a fermentation defoaming control system, method and device based on machine vision. Background Art

[0002] In the industrial fermentation process, defoamers need to be added at different stages of fermentation to defoam. The prior art relies on manual observation of foam through a sight glass or detection by a defoaming electrode. However, the operation of manually observing foam through a sight glass has problems such as a large amount of labor, untimely monitoring, and easy misjudgment; when detecting by a defoaming electrode, on the one hand, the electrode increases the risk of bacterial contamination, and on the other hand, due to the uncertainty of the foam shape, size, density, etc. that may occur when the fermentation broth has different working conditions such as formula and fermentation stage, and the defoaming electrode is limited by its working principle, it may misjudge or even have no signal for foams of different shapes, sizes, and densities, which may lead to foam overflow and easily cause serious consequences such as bacterial contamination and loss of fermentation broth. Summary of the Invention

[0003] The present invention provides a fermentation defoaming control system, method and device based on machine vision to at least partially solve the technical problems of poor monitoring effect and high misjudgment rate in the existing defoaming process.

[0004] The present invention provides a fermentation defoaming control system based on machine vision, and the fermentation defoaming control system includes:

[0005] A fermentation tank, the fermentation tank includes a tank body and a head installed on the top of the tank body, and a sight glass is installed on the head;

[0006] An image acquisition device, the image acquisition device is installed on the sight glass and acquires the current image in the tank through the sight glass;

[0007] A defoamer source, the defoamer source is connected to the fermentation tank through a pipeline, and a switching valve is installed on the pipeline;

[0008] A controller, the controller is used to receive the current image acquired by the image acquisition device, compare the current image with a pre-stored reference image, and generate a comparison result; the controller is also used to generate an opening instruction when the comparison result is that the foam is excessive, and the opening instruction is used to control the switching valve to open.

[0009] In some embodiments, the image acquisition device is a high-definition camera.

[0010] In some embodiments, the controller includes a vision computer and a DCS system.

[0011] The present invention also provides a fermentation defoaming control method based on machine vision, which is applied to the fermentation defoaming control system as described above. The method includes:

[0012] Collect the current image inside the fermentation tank;

[0013] Compare the current image with a pre-stored reference image to generate a comparison result;

[0014] If it is determined that the comparison result indicates excessive foam, generate an opening instruction for controlling the opening of the on-off valve.

[0015] In some embodiments, collecting the current image inside the fermentation tank specifically includes:

[0016] Collect the current image in real time or collect the current image at a target moment at preset time intervals.

[0017] In some embodiments, collecting the current image at a target moment at preset time intervals specifically includes:

[0018] When the fermentation broth in the fermentation tank is in the initial fermentation stage, collect the current image at the target moment at a first preset time interval;

[0019] When the fermentation broth in the fermentation tank is in the stable fermentation stage, collect the current image at the target moment at a second preset time interval;

[0020] Wherein, the first preset time is less than the second preset time.

[0021] In some embodiments, the first preset time is 0.5 - 1.5 minutes, and the second preset time is 4 - 6 minutes.

[0022] In some embodiments, determining that the comparison result indicates excessive foam specifically includes:

[0023] When the similarity between the current image and the pre-stored reference image in the comparison result is greater than or equal to 85%, determine that the comparison result indicates excessive foam.

[0024] In some embodiments, the reference image is a historical sample image pre-stored in the comparison model.

[0025] The present invention also provides a fermentation defoaming control device based on machine vision, which is applied to the fermentation defoaming control system as described above. The device includes:

[0026] An image acquisition unit for collecting the current image inside the fermentation tank;

[0027] A similarity comparison unit for comparing the current image with a pre-stored reference image to generate a comparison result;

[0028] An instruction generation unit, configured to determine that the comparison result is excessive foam, and generate an opening instruction for controlling the opening of the switching valve.

[0029] The fermentation defoaming control system and method based on machine vision provided by the present invention collect the current image in the fermentation tank through an image acquisition device, compare the current image with a pre-stored reference image to generate a comparison result; determine that the comparison result is excessive foam, and generate an opening instruction for controlling the opening of the switching valve. In this way, through image acquisition and comparison of image similarity, and automatically generating an instruction using the comparison result, the automatic control of the switching valve is realized, avoiding the risk of increased bacterial contamination caused by the contact between the traditional defoaming electrode and the fermentation broth, reducing the probability of misjudgment of the foam detection signal, and realizing precise control of fermentation defoaming; solving the technical problems of poor monitoring effect and high misjudgment rate in the existing defoaming process. Description of the Drawings

[0030] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0031] Figure 1 It is a schematic structural diagram of the fermentation defoaming control system based on machine vision provided by the present invention;

[0032] Figure 2 It is a flowchart of the fermentation defoaming control method based on machine vision provided by the present invention;

[0033] Figure 3 It is a structural block diagram of the fermentation defoaming control device based on machine vision provided by the present invention. Detailed Embodiments

[0034] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention with reference to the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0035] To solve the problems existing in the prior art, the present invention provides a fermentation defoaming control system based on machine vision, as Figure 1As shown in the figure, the fermentation defoaming control system includes a fermentation tank 1, an image acquisition device, a defoamer source, and a controller. Among them, the fermentation tank includes a tank body and a head installed on the top of the tank body. A sight glass is installed on the head. The tank body is a place for containing fermentation broth and fermentation reaction. The head is sealed and installed on the top of the tank body to ensure the tightness during the reaction process.

[0036] The image acquisition device is installed on the sight glass and acquires the current image inside the tank through the sight glass. Specifically, the image acquisition device can be a high-definition camera 8 or a single / double binocular camera, etc. Preferably, a high-definition camera 8 is installed on the sight glass of the fermentation tank. During the fermentation and culture process, monitoring and image acquisition are carried out through the high-definition camera 8.

[0037] The defoamer source contains defoamer. The defoamer source is connected to the fermentation tank 1 through a pipeline 2. A switching valve is installed on the pipeline 2. The switching valve includes a main valve 3 and a sub-valve 4. The on / off of the pipeline 2 is realized through the cooperation of the main and sub-valves.

[0038] The controller is used to receive the current image acquired by the image acquisition device, compare the current image with a pre-stored reference image, and generate a comparison result. The controller is also used to generate an opening instruction when the comparison result indicates excessive foam. The opening instruction is used to control the opening of the switching valve. From the perspective of hardware composition, the controller includes a vision computer 7 and a DCS system 6.

[0039] In a specific usage scenario, the fermentation defoaming control system based on machine vision mainly includes a fermentation tank 1, a pipeline 2 for supplementing defoamer, a main valve 3 and a sub-valve 4 for controlling the on / off of the pipeline, a data line / link line 5, a DCS system 6, a machine vision computer 7, a high-definition camera 8, and a sight glass 9. Among them, the fermentation tank 1 is used for culturing fermentation broth. The sight glass 9 is installed on the head of the fermentation tank. The high-definition camera 8 is installed on the sight glass 9 and can monitor the situation of the fermentation broth in the fermentation tank 1 through the sight glass 9. The signal of the high-definition camera 8 is connected to the machine vision computer 7. The machine vision computer 7 is used to compare the current picture taken by the high-definition camera with the reference images of various preset working conditions in real time. When the comparison result with the picture of excessive foam is close, a signal is sent to the DCS system 6. At this time, the DCS system 6 will control the main valve 3 to open, so that the defoamer is added to the fermentation tank 1 through the pipeline 2, thereby achieving the purpose of defoaming.

[0040] In addition to the above system, the present invention also provides a fermentation defoaming control method based on machine vision, which is applied to the fermentation defoaming control system as described above, as Figure 2 shown. The method includes the following steps:

[0041] S110: Acquire the current image inside the fermentation tank; The acquisition of the current image can be real-time acquisition or acquisition at intervals of a preset time.

[0042] S120: Compare the current image with a pre-stored reference image to generate a comparison result. It should be understood that the similarity is compared here, and the comparison result of the similarity can be represented in the form of a percentage. The reference image can be a historical image stored in a database, and the current image can also be input into a comparison model to obtain a comparison result by comparing with the reference image stored in the comparison model. The comparison model can be a model trained based on sample images, working conditions, and corresponding comparison results.

[0043] S130: If it is determined that the comparison result indicates excessive foam, generate an opening instruction, which is used to control the opening of the on-off valve or adjust the opening degree of the on-off valve. For example, when the similarity is higher than a certain preset value, the corresponding comparison result can be determined as excessive foam, and at this time, the on-off valve can be opened or the opening degree of the on-off valve can be increased.

[0044] In this way, the currently taken image is automatically compared with various working-condition foam photos (i.e., reference images) pre-set in the computer. When the comparison result is close to that of the photo with excessive foam, a signal is sent to the DCS system, and the DCS system controls the antifoaming agent valve to add antifoaming agent, thereby inhibiting excessive foam. Since the camera does not come into contact with the fermentation broth, the risk of contamination will not increase due to this. Moreover, as the working time gets longer, the number of photos accumulated by machine vision increases, and the photo comparison model is continuously trained, so that the foam judgment becomes more and more accurate, effectively avoiding losses caused by foam overflow and reducing the risk of contamination caused thereby.

[0045] At different stages of the reaction, the amount of foam generated is different. To ensure the comparison accuracy as much as possible and reduce the calculation amount, the current image at the target moment is collected at intervals of a preset time, specifically including:

[0046] When the fermentation broth in the fermenter is in the initial fermentation stage, the current image at the target moment is collected at intervals of a first preset time;

[0047] When the fermentation broth in the fermenter is in the stable fermentation stage, the current image at the target moment is collected at intervals of a second preset time;

[0048] Among them, the first preset time is less than the second preset time.

[0049] Specifically, the first preset time is 0.5 - 1.5 minutes, and the second preset time is 4 - 6 minutes. That is to say, according to the fermentation formula and process, a video monitoring frequency that conforms to its characteristics is formulated. For example, at the initial stage of fermentation, a photo can be taken every minute, and at the stable fermentation stage, a photo can be taken every 5 minutes.

[0050] In some embodiments, determining that the comparison result is excessive foam specifically includes:

[0051] When the similarity between the current image and a pre - stored reference image is greater than or equal to 85%, determine that the comparison result is excessive foam.

[0052] Specifically, the reference image is a historical sample image pre - stored in the comparison model. The training samples of this comparison model can be foam photos under various working conditions pre - set in the computer. These sample photos include different foam levels, such as no foam, a small amount of foam, medium foam, a large amount of foam, etc. Each sample photo has a label indicating its corresponding foam level. The test samples used during model training are photos of the fermentation broth surface obtained in real - time from the camera, and these photos are used to evaluate the performance of the model in actual applications.

[0053] The training process may include the following steps:

[0054] Data pre - processing: Pre - process the pre - stored sample photos, including operations such as cropping, scaling, and rotation to ensure the consistency of the sample photos with the photos taken in real - time in terms of size and angle. Image enhancement is also required, such as adjusting brightness, contrast, etc., to highlight foam features.

[0055] Feature extraction: Use image - processing techniques to extract key features from the photos; for example, adopt an edge - detection algorithm to identify the contour of the foam.

[0056] Model selection and training: Select a suitable machine - learning model for training. Commonly used models include convolutional neural networks (CNNs). Input the pre - processed photos and their corresponding foam - level labels into the model, and adjust the model parameters through the back - propagation algorithm so that the model can accurately classify the photos into different foam - level categories.

[0057] Model update and optimization: As the working time increases, the system will continuously accumulate new real - time photo data. These data can be used for further training and optimizing the model, enabling the model to adapt to changes under different working conditions and improving the accuracy of foam judgment. For example, online learning or incremental learning methods can be adopted to gradually integrate the new data into the model and update the model parameters to make it more accurately reflect the actual foam situation.

[0058] The fermentation defoaming control system and method based on machine vision avoid the risk of bacterial contamination brought by contact defoaming electrodes, improve the accuracy of foam inspection, and as the use time increases and more process data is accumulated, the recognition and judgment of the high-definition camera will become more accurate, ensuring that the system can timely supplement defoaming agent, avoid liquid escape, reduce the resulting risk of bacterial contamination, and guarantee product quality. Before the system is put into use, pictures of fermentation broth under various working conditions need to be transmitted to the machine vision computer 7 for self-learning. After being put into use, the high-definition camera 8 installed on the sight glass 9 of the fermentation tank 1 conducts real-time video monitoring and takes pictures of the fermentation tank 1, and transmits the pictures to the machine vision computer 7 through the data cable / link cable 5 for comparison with the preset pictures. When the comparison is close to the picture with foam, the machine vision computer 7 sends a signal to the DCS control system 6, and the DCS control system 6 controls the opening of the main defoaming agent valve 3, allowing the defoaming agent to flow into the fermentation tank 1 along the defoaming agent supplement pipeline 2, thereby achieving the purpose of defoaming.

[0059] This system and method not only avoid the risk of increased bacterial contamination due to the contact between the traditional defoaming electrode and the fermentation broth, reduce the probability of misjudgment of the foam detection signal, and achieve precise control of fermentation defoaming. At the same time, due to the self-learning function of the machine vision of the system, the system becomes more and more accurate, and the safety and reliability will be higher and higher.

[0060] The fermentation defoaming control system and method based on machine vision provided by the present invention collect the current image in the fermentation tank through an image acquisition device, compare the current image with a pre-stored reference image to generate a comparison result; if it is determined that the comparison result indicates excessive foam, an opening instruction is generated, and the opening instruction is used to control the opening of the on-off valve. In this way, through image acquisition and comparison of image similarity, and automatically generating an instruction using the comparison result, the automatic control of the on-off valve is realized, avoiding the risk of increased bacterial contamination due to the contact between the traditional defoaming electrode and the fermentation broth, reducing the probability of misjudgment of the foam detection signal, and achieving precise control of fermentation defoaming; solving the technical problems of poor monitoring effect and high misjudgment rate in the existing defoaming process.

[0061] Furthermore, the present invention also provides a fermentation defoaming control device based on machine vision, which is applied to the fermentation defoaming control system as described above. As Figure 3 shown, the device includes:

[0062] An image acquisition unit 100 for collecting the current image in the fermentation tank;

[0063] A similarity comparison unit 200 for comparing the current image with a pre-stored reference image to generate a comparison result;

[0064] An instruction generation unit 300 for determining that the comparison result indicates excessive foam and generating an opening instruction, where the opening instruction is used to control the opening of the on-off valve.

[0065] In some embodiments, collecting the current image inside the fermentation tank specifically includes:

[0066] Collecting the current image in real time or collecting the current image at a target moment at preset time intervals.

[0067] In some embodiments, collecting the current image at a target moment at preset time intervals specifically includes:

[0068] When the fermentation broth in the fermentation tank is in the initial fermentation stage, collecting the current image at the target moment at a first preset time interval;

[0069] When the fermentation broth in the fermentation tank is in the stable fermentation stage, collecting the current image at the target moment at a second preset time interval;

[0070] Wherein, the first preset time is less than the second preset time.

[0071] In some embodiments, the first preset time is 0.5 - 1.5 minutes, and the second preset time is 4 - 6 minutes.

[0072] In some embodiments, determining that the comparison result is excessive foam specifically includes:

[0073] When the similarity between the current image and a pre - stored reference image in the comparison result is greater than or equal to 85%, determining that the comparison result is excessive foam.

[0074] In some embodiments, the reference image is a historical sample image pre - stored in the comparison model.

[0075] The fermentation defoaming control device based on machine vision avoids the risk of bacterial contamination brought by contact - type defoaming electrodes, improves the accuracy of foam inspection, and as the usage time increases and more process data is accumulated, the recognition and judgment of the high - definition camera will be more accurate, ensuring that the system can timely add defoaming agent to avoid liquid escape, reduce the resulting risk of bacterial contamination, and guarantee product quality. Before the system is put into use, pictures of fermentation broth under various working conditions need to be transmitted to the machine vision computer 7 for self - learning. After being put into use, the high - definition camera 8 installed on the sight glass 9 of the fermentation tank 1 conducts real - time video monitoring and takes pictures from the fermentation tank 1, and transmits the pictures to the machine vision computer 7 through the data line / link line 5 for comparison with the preset pictures. When the comparison with the pictures with foam is close, the machine vision computer 7 sends a signal to the DCS control system 6, and the DCS control system 6 controls the opening of the main defoaming agent valve 3, so that the defoaming agent flows into the fermentation tank 1 along the defoaming agent supply pipeline 2, thereby achieving the defoaming purpose.

[0076] This method not only avoids the risk of increased bacterial contamination caused by the contact between the traditional defoaming electrode and the fermentation broth, reduces the probability of misjudgment of the foam detection signal, and realizes precise control of fermentation defoaming. At the same time, due to the self-learning function of the system's machine vision, the system becomes more and more accurate, and the safety and reliability will be higher and higher.

[0077] In the above specific implementation manner, for the fermentation defoaming control device provided by the present invention, the current image in the fermentation tank is collected by the image acquisition device, the current image is compared with the pre-stored reference image to generate a comparison result; if it is determined that the comparison result is excessive foam, an opening instruction is generated, and the opening instruction is used to control the opening of the switching valve. In this way, through image acquisition, comparison of image similarity, and automatic generation of an instruction using the comparison result, automatic control of the switching valve is achieved, avoiding the risk of increased bacterial contamination caused by the contact between the traditional defoaming electrode and the fermentation broth, reducing the probability of misjudgment of the foam detection signal, and realizing precise control of fermentation defoaming; solving the technical problems of poor monitoring effect and high misjudgment rate in the defoaming process in the prior art.

[0078] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

[0079] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. A fermentation defoaming control system based on machine vision, characterized in that: The fermentation and defoaming control system comprises: A fermentation tank, the fermentation tank comprising a tank body and a head mounted on the top of the tank body, wherein a sight glass is mounted on the head; An image acquisition device, which is installed on the viewing mirror and acquires a current image inside the tank through the viewing mirror; A defoamer source, the defoamer source is connected to the fermenter via a pipeline, and a switch valve is installed on the pipeline; A controller is used to receive a current image captured by the image acquisition device, and compare the current image with a pre-stored reference image to generate a comparison result; the controller is also used to generate an opening instruction when the comparison result is excessive foam, and the opening instruction is used to control the opening of the switch valve.

2. The fermentation and defoaming control system based on machine vision according to claim 1, characterized in that: The image acquisition device is a high-definition camera.

3. The fermentation and defoaming control system based on machine vision according to claim 1, characterized in that: The controller includes a visual computer and a DCS system.

4. A fermentation defoaming control method based on machine vision, applied to the fermentation defoaming control system according to any one of claims 1 to 3, characterized in that: The method comprises: Acquire the current image inside the fermentation tank; Comparing the current image with a pre-stored reference image to generate a comparison result; If it is determined that the comparison result is excessive foam, an opening instruction is generated, and the opening instruction is used to control the switch valve to open.

5. The fermentation defoaming control method based on machine vision according to claim 4, characterized in that: Collect the current image inside the fermentation tank, including: The current image is acquired in real time or the current image at a target moment is acquired at a preset interval.

6. The fermentation defoaming control method based on machine vision according to claim 5, characterized in that: The current image at the target time is collected at preset intervals, including: When the fermentation liquid in the fermentation tank is in the initial fermentation stage, collecting the current image at the target moment at intervals of a first preset time; When the fermentation liquid in the fermentation tank is in a stable fermentation stage, collecting the current image at the target moment at intervals of a second preset time; Wherein, the first preset time is shorter than the second preset time.

7. The fermentation defoaming control method based on machine vision according to claim 6, characterized in that: The first preset time is 0.5-1.5 minutes, and the second preset time is 4-6 minutes.

8. The fermentation defoaming control method based on machine vision according to claim 6, characterized in that: Determining that the comparison result is excessive foam specifically includes: When the comparison result is that the similarity between the current image and the pre-stored reference image is greater than or equal to 85%, it is determined that the comparison result is excessive foam.

9. The fermentation defoaming control method based on machine vision according to claim 6, characterized in that: The reference image is a historical sample image pre-stored in the comparison model.

10. A fermentation defoaming control device based on machine vision, applied to the fermentation defoaming control system according to any one of claims 1 to 3, characterized in that: The device comprises: An image acquisition unit, used for acquiring a current image in the fermentation tank; A similarity comparison unit, used to compare the current image with a pre-stored reference image to generate a comparison result; The instruction generating unit is used to determine that the comparison result is excessive foam, and then generate an opening instruction, wherein the opening instruction is used to control the opening of the switch valve.

Images