Waste water-based drilling fluid flocculation image recognition method and device based on machine vision

Abstract

The invention discloses a waste water-based drilling fluid flocculation image recognition method and device based on machine vision. The method comprises the steps that a flocculation video, shot in real time, of waste water-based drilling fluid is obtained; transmitting the flocculation video to a server for decoding to obtain a flocculation image containing a flocculation reaction in the flocculation video; inputting the flocculation image into a preset recognition model obtained by training, and recognizing to obtain a flocculation reaction degree value, so as to determine whether a flocculating agent is continuously added or not according to the flocculation reaction degree value; wherein the preset recognition model is obtained based on flocculation reaction image training in different stages. The preset recognition model is trained in advance based on the flocculation reaction images at different stages, and different flocculation reaction degree values generated after the flocculant is added into the waste water-based drilling fluid can be recognized, so that the flocculation condition is effectively recognized on line, and manual treatment is not needed.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of machine vision recognition, and in particular to a method and device for recognizing flocculation images of waste water-based drilling fluid based on machine vision. Background Art

[0002] Water-based drilling fluid is a multiphase colloid-suspension system that can circulate in the drilling pipeline and is composed of water, drilling fluid additives, preservatives, pH regulators, suspending agents, dispersants, viscosity enhancers, surfactants and other chemical substances. Its function is to lubricate the drill bit, stabilize the well wall and carry drill cuttings to ensure safe and efficient drilling operations. Drilling is affected by different environments. In order to meet the needs of different drilling, it is necessary to design a variety of drilling fluids, among which water-based drilling fluids have a wider range of applications.

[0003] When treating waste water-based drilling fluid flocculation, the flocculants used include inorganic flocculants, organic polymer flocculants, microbial flocculants, composite flocculants, etc. The flocculation mechanisms of these flocculants include electrical neutralization mechanism, adsorption bridging mechanism, net capture and sweeping mechanism, etc. In the flocculation process, the existing technology is generally based on human eye visual inspection or mathematical model method, light transmission pulsation detection technology, etc. to identify whether the water-based drilling fluid is completely flocculated.

[0004] However, the flocculation reaction between the flocculant and the waste drilling fluid is short-lived, generally completed within 30 seconds, followed by a diffusion process. It is difficult to accurately identify whether the water-based drilling fluid has been completely flocculated in a short period of time. In addition, there are a large number of rock cuttings, silt, and impurities floating on the liquid surface in the waste water-based drilling fluid that do not participate in the flocculation reaction, which affects the flocculation effect of the waste water-based drilling fluid. As a result, the existing technology has the following problems in determining whether the waste water-based drilling fluid is completely flocculated:

[0005] Manual observation has a lag, and labor costs will greatly increase expenses, resulting in very low efficiency of visual inspection by the human eye; the mathematical model method requires the construction of a specific mathematical model for a specific environment, taking waste water quality parameters (such as pH value, temperature, turbidity, etc.) as variables to establish a mathematical model of the functional relationship between it and the amount of flocculant used. This method lacks autonomous learning and adaptability, has high requirements on the accuracy of the detection instrument, and does not have generalization capabilities; the light transmittance pulsation detection technology uses the fluctuation state of the light transmittance intensity through the flocculant to calculate the particle changes of the flocs formed after coagulation. In the automatic dosing control system, the influence of factors such as water quality and water volume on the treatment effect can be eliminated in time, and the target value of the turbidity of the water after sedimentation is basically stable, which has good economic and social benefits. However, for waste water-based drilling fluid, the waste liquid contains a large amount of sediment, rock chips and other substances, the turbidity is too high, and the light transmittance is extremely low, so it cannot be applied. Summary of the invention

[0006] In view of the above problems, embodiments of the present invention are proposed to provide a method and device for machine vision-based waste water-based drilling fluid flocculation image recognition that overcomes the above problems or at least partially solves the above problems.

[0007] According to one aspect of an embodiment of the present invention, a method for identifying flocculation images of waste water-based drilling fluid based on machine vision is provided, the method comprising:

[0008] Get real-time video of flocculation of waste water-based drilling fluids;

[0009] The flocculation video is transmitted to the server for decoding and processing, and a flocculation image containing a flocculation reaction in the flocculation video is obtained;

[0010] The flocculation image is input into the trained preset recognition model to identify the flocculation reaction degree value, so as to determine whether to continue adding flocculant according to the flocculation reaction degree value; wherein the preset recognition model is trained based on flocculation reaction images at different stages.

[0011] According to another aspect of an embodiment of the present invention, there is provided a device for identifying flocculation images of waste water-based drilling fluid based on machine vision, which includes:

[0012] A real-time acquisition module, adapted to acquire a real-time video of flocculation of waste water-based drilling fluid;

[0013] An image processing module is adapted to transmit the flocculation video to the server for decoding and processing, and obtain a flocculation image containing a flocculation reaction in the flocculation video;

[0014] The recognition module is suitable for inputting the flocculation image into the trained preset recognition model to recognize the flocculation reaction degree value, so as to determine whether to continue adding flocculant according to the flocculation reaction degree value; wherein the preset recognition model is trained based on flocculation reaction images at different stages.

[0015] According to another aspect of an embodiment of the present invention, there is provided a computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0016] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute operations corresponding to the above-mentioned method for identifying flocculation images of waste water-based drilling fluid based on machine vision.

[0017] According to another aspect of an embodiment of the present invention, a computer storage medium is provided, wherein at least one executable instruction is stored in the storage medium, and the executable instruction enables a processor to perform operations corresponding to the above-mentioned method for identifying flocculation images of waste water-based drilling fluid based on machine vision.

[0018] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising at least one executable instruction, wherein the executable instruction enables a processor to perform operations corresponding to the above-mentioned method for identifying flocculation images of waste water-based drilling fluid based on machine vision.

[0019] According to the machine vision-based waste water-based drilling fluid flocculation image recognition method and device provided in an embodiment of the present invention, a preset recognition model is pre-trained based on flocculation reaction images at different stages, which can identify different flocculation reaction degree values ​​produced after adding flocculants to the waste water-based drilling fluid, thereby realizing online effective identification of flocculation conditions without manual processing.

[0020] The above description is only an overview of the technical solution of the embodiment of the present invention. In order to more clearly understand the technical means of the embodiment of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present invention more obvious and easy to understand, the specific implementation method of the embodiment of the present invention is specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the embodiments of the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0022] Figure 1 A flow chart of a method for identifying flocculation images of waste water-based drilling fluid based on machine vision according to an embodiment of the present invention is shown;

[0023] Figure 2 The structural schematic diagram of the flocculation bin mode is shown;

[0024] Figure 3 A schematic diagram of the structure of the beaker mode is shown;

[0025] Figure 4 A schematic diagram of the structure of an image acquisition system is shown;

[0026] Figure 5 The schematic diagram of the structure of the real-time collection flocculation tank is shown;

[0027] Figure 6 A schematic diagram of the structure of a device for identifying flocculation images of waste water-based drilling fluid based on machine vision according to an embodiment of the present invention is shown;

[0028] Figure 7 A schematic diagram of the structure of a computing device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0029] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0030] Figure 1 FIG. 4 is a flowchart of a method for identifying flocculation images of waste water-based drilling fluid based on machine vision according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0031] Step S101, obtaining a real-time video of flocculation of waste water-based drilling fluid.

[0032] Add the waste water-based drilling fluid to the flocculation tank, use the flocculant dosing device to continuously add flocculants to the flocculation tank of the waste water-based drilling fluid, and use a stirrer to stir it. At the same time, the flocculation reaction of the waste water-based drilling fluid with continuous flocculants added is filmed in real time to obtain a real-time flocculation video. The obtained flocculation video can be transmitted to the image acquisition and processing software on the server side through a network cable for processing.

[0033] When shooting, you can use an industrial CMOS camera with a unique POE interface to shoot the waste water-based drilling fluid with added flocculants to obtain a flocculation video. When transmitting, it can be transmitted through a Gigabit Ethernet cable. The flocculation reaction time takes a certain amount of time, such as completing the flocculation reaction within 30 seconds. Shooting a flocculation video can record all images of the entire flocculation process at different time periods from the addition of flocculants to the beginning of flocculation to complete flocculation.

[0034] When shooting, an observation window can be set in the flocculation tank, and the observation window can be set as an acrylic transparent box. A light source can be placed behind the camera, and a reflector can be set behind the light source. The light absorption device is set in front of the observation window, such as Figure 5 As shown, when the light source irradiates the flocs in the flocculation tank, the light will be reflected to the camera. When the light source does not irradiate the flocs, the light passes through the glass plate to the light absorbing device and is completely absorbed. The bright spots of each frame image in the flocculation video captured by the camera are the flocs, thereby ensuring clarity and contrast, which is conducive to subsequent identification. The light absorbing device can be set with existing light absorbing materials, which is not limited here.

[0035] Step S102: transmitting the flocculated video to the server for decoding, and obtaining a flocculated image containing a flocculation reaction in the flocculated video.

[0036] The flocculated video can be transmitted through a Gigabit Ethernet cable. Compared with the existing image transmission method of transmitting the image through a cable to an image acquisition card, the computer acquires the image in the acquisition card, and stores it in the computer for processing, the flocculated video can be directly stored in the background server without the need for an acquisition card, which improves processing efficiency.

[0037] After the flocculation video is transmitted to the backend server, the flocculation video can be decoded and decomposed to obtain a flocculation image containing a flocculation reaction, so as to perform online recognition based on the flocculation image. The flocculation video can be processed by, for example, a visual sensor based on a standard bus to input the flocculation video to a designated image acquisition and image processing software to obtain a flocculation image for recognition.

[0038] Step S103, inputting the flocculation image into the trained preset recognition model to identify and obtain a flocculation reaction degree value, so as to determine whether to continue adding flocculants according to the flocculation reaction degree value.

[0039] To facilitate online recognition, this embodiment pre-trains to obtain a preset recognition model, which adopts a target detection model, and is trained by updating the network weights of the preset recognition model based on a back-propagation algorithm. The target detection model can adopt version 6.1 of the YOLOv5 algorithm, and completes training through data enhancement, feature extraction, feature fusion, target detection, classification and regression, and model optimization. During training, learning rate scheduling strategies (Cosine Annealing) and weight decay (Weight Decay) can be used to improve model performance.

[0040] Specifically, in this embodiment, YOLOv5 6.1 adopts CSPDarknet as the Backbone backbone feature extraction network. CSPDarknet introduces the CSPNet (Cross Stage Hierarchical Network) structure based on Darknet53. The YOLOv5 backbone feature extraction network adopts the C3 structure, which has a large number of parameters, slow detection speed, and limited application. It is difficult to be applied to scenes requiring low latency or fast response speed. In this embodiment, the backbone feature extraction network is replaced with a PP-LCNet network to achieve lightweight network model and balance speed and accuracy. In terms of data enhancement, in this embodiment, YOLOv5 6.1 improves the Mosaic data enhancement method, adopts more image cropping methods, improves the effect of data enhancement, and further introduces the Random Grid Shuffle method to further enhance the generalization ability of the model and make the algorithm more widely applicable. In terms of loss function, in this embodiment, YOLOv5 6.1 adopts GIoU loss (Generalized IoULoss) as the main loss function. On the basis of GIoU, the loss of the detection box scale is increased, the loss of length and width is increased, and the center point offset, length and width ratio difference, and area difference of the image bounding box are further considered. Compared with the traditional IoU that only considers the intersection and union of the bounding boxes, CloU also considers the mismatch of the bounding box when calculating the intersection, making the predicted box more consistent with the real box, solving the problem of data sample imbalance, and accelerating the convergence of the network.

[0041] Furthermore, in this embodiment, YOLOv5 6.1 optimizes the neck network, adopts the SPP (Spatial PyramidPooling) layer and the PAN (Path Aggregation Network) layer to improve the neck network, and the improved module is the Eff-QAFPN structure. The Eff-QAFPN structure can improve the speed and efficiency of the network by reducing the amount of calculation and memory usage, and perform target detection more efficiently. In this embodiment, Decoupled Head is used to replace the original detection head network to extract image features and predict the category of each pixel. Traditional image segmentation networks usually integrate feature extraction and pixel prediction processes in the same network. Decoupled Head decouples these two processes and processes them separately, which can better handle semantic information of different scales and fineness. The Decoupled Head network adds one or more branches to the feature maps of different levels of the backbone network, integrates feature information from different scales, and improves the segmentation ability of multi-scale targets. By separating pixel-level prediction from feature extraction, the semantic information between low-level and high-level features can be better utilized. Processing pixel-level prediction as an independent task can better preserve details and edge information, thereby improving segmentation accuracy and detail retention. Furthermore, the Decoupled Head network can be expanded and modified as needed, such as adding more branches or adjusting the structure of the branches to adapt to different tasks and data set requirements.

[0042] During training, the flocculation reaction images at different time periods after adding the flocculant are collected based on the flocculant image data collection system. The flocculant image data collection system includes: a flocculation processing system and an image collection system. In this embodiment, the flocculation processing system is used to collect flocculation reaction images to train the preset recognition model. The flocculation processing system can use the flocculation bin mode to obtain the flocculation reaction image, or it can use the beaker mode. Specifically, the flocculation bin mode is as follows: Figure 2 As shown, it includes a flocculation bin 6, a filter press butterfly valve 4 and a sewage butterfly valve 5 are arranged in the flocculation bin 6, an agitator 7 connected to the flocculation bin 6, an observation pipeline butterfly valve 3 connected to the flocculation bin 6, an observation pipeline 2 connected to the observation pipeline butterfly valve 3, and a single screw circulation pump 1 connected to the observation pipeline 2 and the filter press butterfly valve 4. The single screw circulation pump 1 can be used, for example, with a flow rate of 5m 3 / h, pressure 0.6Mpa, power 3kw, speed 960r / min, G30-1 type single screw pump is used as the observation pipeline drilling fluid circulation pump. The agitator 7 can use the BLD11-17-3KW type agitator with a speed of 82-8r / min, coupling connection, shaft length of 3 meters, shaft diameter of 38mm, impeller diameter of 500mm, and two layers of double blades. The flocculation chamber 6 can process 1 cubic meter of drilling fluid. In order to better collect images during the flocculation process, on one side of the flocculation chamber 6 (such as Figure 2 Two channels are set at different heights (as shown on the left side), which are respectively connected to two observation pipes 2 at different heights, and the switch is controlled by the observation pipe butterfly valve 3. Since the waste water-based drilling fluid contains a large amount of rock cuttings, chemical substances that do not participate in the reaction, etc., in order to prevent the above substances from accumulating in the observation pipe 2, one side of the bottom of the flocculation bin 6 can be raised, such as at an angle of 7 degrees, to facilitate subsequent cleaning. The material of the observation pipe 2 can be selected as tempered glass, and a nano-coating with oil-proof and anti-corrosion functions is sprayed on the inside, which is oil-proof and anti-corrosion while also convenient for observation. The beaker mode can be as follows Figure 3 As shown, it includes a glass cup 13 filled with waste water-based drilling fluid and flocculant and a beaker stirrer 14. Among them, the glass cup 13 is made of acrylic glass, and the beaker stirrer 14 is a beaker stirrer with adjustable drilling speed.

[0043] The above is an example, which is set according to the implementation situation and is not limited here. By constructing the above flocculation treatment system, the flocculation reaction of adding flocculants to the water-based drilling fluid can be obtained in real time.

[0044] The image acquisition system includes Figure 4 The camera 9 shown, the camera bracket 8 supporting the camera 9, the high-brightness LED flash-free explosion-proof lamp 12, the network cable 11 for transmitting the image, the computer 10 for collecting the image, and the like Figure 3 The switch 15 shown. The camera 9 can be a Basler camera with a pixel of 20.2MP, a frame rate of 27fps, explosion-proof and corrosion-proof functions, which supports image compression, modification of corresponding algorithms, increase or decrease of camera functions, and adjustable frame rate and resolution. Figure 2 The structural schematic diagram of the flocculation bin mode is shown and will not be described in detail here.

[0045] By using the floc image data acquisition system, the waste water-based drilling fluid of different well depths, such as the water-based drilling fluid system BIODRILLA, can be flocculated, and images from before the start of flocculation to after the flocculation reaction is complete can be collected. The flocculation reaction takes a certain amount of time, such as 30 seconds. The camera exposure time can be set to expose once every 5 seconds to capture the flocculation reaction image. Specifically, taking the flocculation bin mode as an example, the waste water-based drilling fluid is stirred evenly and then sampled, and flocculants are added to the flocculation bin 6. The stirrer 7 slowly stirs evenly, and the camera 9 collects images. Through the network cable 11 and the switch 15, the image is transmitted to the computer 10 based on Gigabit Ethernet. For the beaker mode, flocculant can be added to the waste water-based drilling fluid, and the camera 9 can be turned on before the beaker stirrer 14 is used for stirring. The camera focal length and aperture size are adjusted, and the camera 9 is placed at a distance of 70 cm from the glass 13. The camera 9 is placed on the side with a light source, and the light source is a high-brightness LED flash-free explosion-proof lamp 12. The camera 9 is set to expose once every 5 seconds, etc., to ensure that the floc morphology on the image taken will not be repeated in large quantities until the flocculation process is completed.

[0046] Based on the above flocculation bin mode, beaker mode, etc., flocculation reaction images of waste water-based drilling fluids at different well depths and different contents of flocculants can be simulated and collected. After obtaining flocculation reaction images of different time periods, the water-based drilling fluids can be classified according to the different colors presented in different flocculation reactions, and the flocculation reaction images can be annotated based on the classification to obtain sample data. During the annotation process, all flocculation reaction images are classified according to the colors presented during the flocculation reaction. For example, after the flocculation reaction part in the image is classified by color through a rectangular frame, the corresponding category is annotated, that is, the flocculation reaction degree value. After obtaining the sample data, the sample data can also be divided into a training set, a validation set, and a test set according to the proportion to verify the accuracy of the preset recognition model training. The sample data is input into the preset recognition model for training. For example, yolov5 6.1 is used to build the preset recognition model, and the predicted flocculation reaction degree value is output. It is compared with the flocculation reaction degree value of the labeled sample data, and the back propagation algorithm is used to update the network weight. The loss function reduces the loss value. The training obtains a preset recognition model for identifying the flocculation reaction effect of abandoned water-based drilling fluid, in which the prediction result and the sample data basically match.

[0047] The sample data is enhanced in the preset recognition model, such as uniformly scaling the input sample data to 608×608 size, and processing the image through the Mosaic enhancement method, including randomly cropping, scaling, rotating and other operations on multiple images, and then splicing them into a new image to increase the diversity of the sample data and improve the generalization ability of the model. The enhanced image is input into the Backbone network to extract multi-scale features. The extracted feature map is input into the Neck network, and the multi-scale feature fusion is performed using the Feature Pyramid Network (FPN). The FPN structure can fuse feature maps of different scales from top to bottom, so that the model can better capture targets of different sizes. On the feature map generated by the Neck network, target detection is performed. Target detection includes two parts: one is to generate prediction boxes (bounding box proposals), and the other is to classify and regress the prediction boxes. The prediction box is generated using the Spatial Pyramid Pooling (SPP) layer and the Head output layer unique to YOLOv5. The Head output layer contains multiple branches, each of which is responsible for predicting targets of different scales. The loss function uses GIoU (Generalized IoU) loss, and combines diou (Distance IoU) to perform non-maximum suppression (NMS) to eliminate redundant prediction boxes. The generated prediction boxes are classified and regressed. The classification uses the softmax layer, and the regression uses the smooth L1 loss (smooth L1 loss) and cluster loss (cluster loss), etc., and finally the trained preset recognition model is obtained.

[0048] After obtaining the trained preset recognition model, the flocculation effect of the waste drilling fluid can be identified online based on the preset recognition model, such as inputting the flocculation image into the trained preset recognition model to identify the flocculation reaction degree value. The flocculation reaction degree value is judged to determine whether it is a complete flocculation reaction value; if not, the flocculant dosing device continues to add flocculant until the flocculation reaction degree value obtained by the preset recognition model is a complete flocculation reaction value, thereby realizing automatic addition of flocculant. Compared with the transmittance pulsation method, it is not affected by turbidity and transmittance, and the flocculation accuracy is further improved. Furthermore, when the flocculation reaction degree value reaches the complete flocculation reaction value, the addition of flocculant can be stopped, and the solid-liquid separation can be performed by a filter press to obtain the treated drilling fluid and mud cake, which further improves the existing water-based drilling fluid reduction and reuse process. The entire process does not require automated processing, which greatly reduces labor costs.

[0049] According to the machine vision-based waste water-based drilling fluid flocculation image recognition method provided by an embodiment of the present invention, a preset recognition model is pre-trained based on flocculation reaction images at different stages, which can identify different flocculation reaction degree values ​​generated after adding flocculants to the waste water-based drilling fluid, thereby realizing online effective identification of flocculation conditions without manual processing.

[0050] Figure 6 FIG. 1 is a schematic diagram showing the structure of a waste water-based drilling fluid flocculation image recognition device based on machine vision provided by an embodiment of the present invention. Figure 6 As shown, the device comprises:

[0051] A real-time acquisition module 610, adapted to acquire a real-time video of flocculation of the waste water-based drilling fluid;

[0052] The image processing module 620 is adapted to transmit the flocculation video to the server for decoding and processing, and obtain a flocculation image containing a flocculation reaction in the flocculation video;

[0053] The recognition module 630 is adapted to input the flocculation image into a trained preset recognition model to recognize the flocculation reaction degree value, so as to determine whether to continue adding flocculants according to the flocculation reaction degree value; wherein the preset recognition model is trained based on flocculation reaction images at different stages.

[0054] Optionally, the device also includes: a training module 640, which is suitable for collecting flocculation reaction images in different time periods after adding flocculants based on the flocculant image data acquisition system; classifying the water-based drilling fluid according to the different colors presented in different flocculation reactions, and annotating the flocculation reaction images based on the classification to obtain sample data; training the preset recognition model according to the sample data to obtain the trained preset recognition model.

[0055] Optionally, the floc image data acquisition system includes: a flocculation processing system and an image acquisition system;

[0056] The flocculation treatment system includes a flocculation bin, a filter press butterfly valve and a sewage butterfly valve provided in the flocculation bin, an agitator connected to the flocculation bin, an observation pipeline butterfly valve connected to the flocculation bin, an observation pipeline connected to the observation pipeline butterfly valve, and a single screw circulation pump connected to the observation pipeline and the filter press butterfly valve; or a glass cup filled with waste water-based drilling fluid and flocculant and a beaker agitator;

[0057] The image acquisition system includes a camera, a camera bracket supporting the camera, a high-brightness LED flash-free explosion-proof lamp, a network cable for transmitting images, a computer for acquiring images, and a switch.

[0058] Optionally, the preset recognition model adopts a target detection model, and the network weights of the preset recognition model are updated for training based on the back propagation algorithm; the target detection model is version 6.1 of the yolov5 algorithm, the loss function of the target detection model is the GIoU loss function, the neck network of the target detection model is the Eff-QAFPN structure, and the head network is the Decoupled Head network.

[0059] Optionally, the real-time acquisition module 610 is further adapted to:

[0060] Add flocculant into the flocculation tank containing the waste water-based drilling fluid for stirring; the flocculation tank is provided with an observation window; the observation window is an acrylic transparent box;

[0061] An industrial CMOS camera is used to obtain real-time flocculation video; a light source is arranged on the rear side of the industrial CMOS camera and a reflector is arranged behind the light source; and a light absorbing device is arranged in front of the observation window.

[0062] Optionally, the identification module 630 is further adapted to:

[0063] The flocculation image is input into the trained preset recognition model to identify and obtain the flocculation reaction degree value;

[0064] Determine whether the flocculation reaction degree value is the flocculation reaction complete value;

[0065] If not, continue to add flocculant until the flocculation reaction degree value reaches the flocculation reaction complete value.

[0066] The description of each module above refers to the corresponding description in the method embodiment and will not be repeated here.

[0067] An embodiment of the present invention further provides a non-volatile computer storage medium storing at least one executable instruction, which can execute operations corresponding to the machine vision-based waste water-based drilling fluid flocculation image recognition method in any of the above method embodiments.

[0068] An embodiment of the present application provides a computer program product, which includes at least one executable instruction or computer program, which enables a processor to perform operations corresponding to the machine vision-based waste water-based drilling fluid flocculation image recognition method in any of the above-mentioned method embodiments.

[0069] Figure 7 A schematic diagram of the structure of a computing device according to an embodiment of the present invention is shown, and the specific implementation of the embodiment of the present invention does not limit the specific implementation of the computing device.

[0070] like Figure 7As shown, the computing device may include: a processor 702 , a communication interface 704 , a memory 706 , and a communication bus 708 .

[0071] in:

[0072] The processor 702 , the communication interface 704 , and the memory 706 communicate with each other via a communication bus 708 .

[0073] The communication interface 704 is used to communicate with other devices such as clients or other servers.

[0074] The processor 702 is used to execute the program 710, and specifically can execute the relevant steps in the above-mentioned embodiment of the method for identifying flocculation images of waste water-based drilling fluid based on machine vision.

[0075] Specifically, the program 710 may include program codes, which include computer operation instructions.

[0076] The processor 702 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computing device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0077] The memory 706 is used to store the program 710. The memory 706 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0078] Program 710 can be specifically used to enable processor 702 to execute the waste water-based drilling fluid flocculation image recognition method based on machine vision in any of the above-mentioned method embodiments. The specific implementation of each step in program 710 can refer to the corresponding descriptions in the corresponding steps and units in the above-mentioned waste water-based drilling fluid flocculation image recognition embodiment based on machine vision, which will not be repeated here. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described equipment and modules can refer to the corresponding process description in the above-mentioned method embodiment, which will not be repeated here.

[0079] The algorithm or display provided herein is not inherently related to any particular computer, virtual system or other device. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious that the structure required for constructing such systems. In addition, the embodiment of the present invention is not directed to any specific programming language either. It should be understood that various programming languages ​​can be utilized to realize the content of the embodiment of the present invention described herein, and the description of the above specific language is to disclose the preferred implementation of the embodiment of the present invention.

[0080] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0081] Similarly, it should be understood that in order to streamline the embodiments of the present invention and to aid in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be interpreted as reflecting the following intention: that the claimed embodiments of the present invention require more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, the inventive aspects lie in less than all the features of the single embodiment disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present invention.

[0082] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0083] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0084] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The embodiments of the present invention may also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing an embodiment of the present invention may be stored on a computer-readable medium, or may have the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0085] It should be noted that the above embodiments illustrate rather than limit the present invention, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim listing a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.

Claims

1. A method for identifying flocculation images of waste water-based drilling fluid based on machine vision, characterized in that: Methods include: Get real-time video of flocculation of waste water-based drilling fluids; Transmitting the flocculation video to a server for decoding to obtain a flocculation image containing a flocculation reaction in the flocculation video; The flocculation image is input into a trained preset recognition model to identify a flocculation reaction degree value, so as to determine whether to continue adding flocculant according to the flocculation reaction degree value; wherein the preset recognition model is trained based on flocculation reaction images at different stages.

2. The method according to claim 1, characterized in that The method further comprises: Based on the flocculent image data acquisition system, collect the flocculation reaction images at different time periods after adding flocculants; Classify the water-based drilling fluid according to the different colors presented in different flocculation reactions, and annotate the flocculation reaction image based on the classification to obtain sample data; The preset recognition model is trained according to the sample data to obtain a trained preset recognition model.

3. The method according to claim 2, wherein: The floc image data acquisition system comprises: a flocculation processing system and an image acquisition system; The flocculation treatment system includes a flocculation bin, a filter press butterfly valve and a sewage butterfly valve are arranged in the flocculation bin, an agitator connected to the flocculation bin, an observation pipeline butterfly valve connected to the flocculation bin, an observation pipeline connected to the observation pipeline butterfly valve, and a single screw circulation pump connected to the observation pipeline and the filter press butterfly valve; or a glass cup filled with waste water-based drilling fluid and flocculant and a beaker agitator; The image acquisition system includes a camera, a camera bracket supporting the camera, a high-brightness LED non-strobe explosion-proof lamp, a network cable for transmitting images, a computer for acquiring images, and a switch.

4. The method according to claim 1, characterized in that: The preset recognition model adopts a target detection model, and the network weights of the preset recognition model are updated for training based on the back propagation algorithm; the target detection model is version 6.1 of the yolov5 algorithm, the loss function of the target detection model is the GIoU loss function, the neck network of the target detection model is the Eff-QAFPN structure, and the head network is the Decoupled Head network.

5. The method according to claim 1, characterized in that The obtaining of the real-time flocculation video of the waste water-based drilling fluid further comprises: Adding flocculant into a flocculation tank containing waste water-based drilling fluid for stirring; the flocculation tank is provided with an observation window; the observation window is an acrylic transparent box; An industrial CMOS camera is used to obtain a real-time flocculation video; wherein a light source is arranged on the rear side of the industrial CMOS camera and a reflector is arranged behind the light source; and a light absorbing device is arranged in front of the observation window.

6. The method according to claim 1, characterized in that The step of inputting the flocculation image into a preset recognition model obtained through training to obtain a flocculation reaction degree value through recognition, and determining whether to continue adding flocculants according to the flocculation reaction degree value further includes: Inputting the flocculation image into a trained preset recognition model to identify and obtain a flocculation reaction degree value; Determining whether the flocculation reaction degree value is a flocculation reaction complete value; If not, continue to add flocculant until the flocculation reaction degree value reaches the flocculation reaction complete value.

7. A waste water-based drilling fluid flocculation image recognition device based on machine vision, characterized in that: The device includes: A real-time acquisition module, adapted to acquire a real-time video of flocculation of waste water-based drilling fluid; An image processing module, adapted to transmit the flocculation video to a server for decoding and processing, and obtain a flocculation image containing a flocculation reaction in the flocculation video; The recognition module is suitable for inputting the flocculation image into a trained preset recognition model to recognize the flocculation reaction degree value, so as to determine whether to continue adding flocculant according to the flocculation reaction degree value; wherein the preset recognition model is trained based on flocculation reaction images at different stages.

8. A computing device, characterized in that include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the method for machine vision-based waste water-based drilling fluid flocculation image recognition as described in any one of claims 1-6.

9. A computer storage medium, characterized in that The storage medium stores at least one executable instruction, and the executable instruction enables the processor to execute operations corresponding to the method for machine vision-based waste water-based drilling fluid flocculation image recognition as described in any one of claims 1 to 6.

10. A computer program product, comprising at least one executable instruction, wherein the executable instruction enables a processor to execute operations corresponding to the method for identifying flocculation images of waste water-based drilling fluid based on machine vision as described in any one of claims 1 to 6.

Images