Image-based liquid level state information monitoring method for constant-level oil cup with limiting glass
Through the image-based monitoring method, the liquid level information of the glass constant oil cup is identified by using the YOLO target detection network, which solves the problem of inefficient traditional manual monitoring and achieves efficient and accurate liquid level status monitoring.
Patent Information
- Application Number
- CN202411652108.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-19
- Publication Date
- 2025-02-18
AI Technical Summary
The liquid level control of traditional glass constant oil cups relies on manual monitoring, and there are errors, which are not suitable for the requirements of modern industry for high efficiency and high precision, especially in complex forms and complex light environments with limit glass constant oil cups. , the accuracy of image recognition is challenged.
Image-based monitoring method is adopted, by acquiring and labeling sample images, the YOLO target detection network is trained, the outer contour and liquid level line of the oil cup are identified, and the liquid level status information is calculated based on the instrument prior information of the inspection point.
It realizes automatic identification, improves the efficiency and accuracy of liquid level status information monitoring, and is suitable for various types of constant-level oil cups with limit glass, with wide application prospects.
Smart Images
Figure CN119478835A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present disclosure relate to the field of industrial automation detection technology, and in particular to an image-based method for monitoring liquid level status information of a constant-position oil cup with a limit glass. Background Art
[0002] In the field of petroleum refining and related industries, comprehensive liquid level control is a key link to ensure production safety and stable operation of equipment. As a device for maintaining a constant liquid level, the glass constant level oil cup is widely used in various industrial equipment to ensure the stability of the oil volume and the normal operation of the equipment. Traditional glass constant level oil cup liquid level control mostly relies on manual monitoring, which has certain errors and is not suitable for the high efficiency and high precision requirements of modern industry. With the development of artificial intelligence and image processing technology, automatic recognition of liquid level status information through image processing technologies such as computer vision can significantly improve the accuracy and real-time performance of liquid level monitoring. However, the complex shape of the glass constant level oil cup with limit and the complex light changes in the on-site environment pose challenges to the accuracy of image recognition. Traditional image processing methods often perform poorly under these conditions.
[0003] Therefore, it is necessary to improve one or more problems existing in the above-mentioned related technical solutions.
[0004] It should be noted that this section is intended to provide background or context for the technical solutions of the present disclosure stated in the claims. The description herein is not admitted to be prior art by virtue of being included in this section. Summary of the invention
[0005] The purpose of the embodiments of the present disclosure is to provide an image-based method for monitoring the liquid level status information of a constant-level oil cup with a limit glass, thereby overcoming one or more problems caused by the limitations and defects of the relevant technology at least to a certain extent.
[0006] According to an embodiment of the present disclosure, a method for monitoring the liquid level status information of a constant-position oil cup with a limit glass based on an image is provided, the method comprising: Acquire a sample image, and mark the outer contour of the oil cup and the liquid level line area of the constant-level oil cup with limiting glass in the sample image; Input the labeled sample image into the YOLO target detection network for training to obtain a trained target detection model; Using a camera to collect a large field of view image containing the constant position oil cup with limit glass to be identified at a preset inspection point; Using the trained target detection model to identify the large field of view image, obtaining first outer contour position information of the constant position oil cup with limit glass; Adjust the angle of the camera according to the first outer contour position information and re-capture a small field of view image including the constant position oil cup with limit glass to be identified; Using the trained target detection model to identify the small field of view image, obtain the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass; The oil cup liquid level status information is obtained according to the instrument priori information, the second outer contour position information and the liquid level line position information.
[0007] Furthermore, the step of acquiring a sample image and marking the outer contour of the oil cup and the liquid level line area of the constant-position oil cup with a limiting glass in the sample image includes: Acquire the sample image, and perform data cleaning on the sample image to remove duplicate images, missing images and useless images; The outer contour of the oil cup and the liquid level line area of the sample image after cleaning are marked.
[0008] Furthermore, the method also includes: Constructing a YOLO target detection network; wherein the YOLO target detection network includes an input end, a backbone, a neck, and a prediction output end; Performing image preprocessing on the input end of the YOLO target detection network for model input, wherein the image preprocessing includes image size adjustment, RGB value normalization of image pixels, and image data enhancement; The Backbone is Cross Stage Partial Darknet 53; The Neck includes a Path Aggregation Network and a Feature Pyramid Network. The prediction output end includes a detection head, which includes a convolution layer, a candidate box generation layer, a prediction layer and an activation function. The prediction output end is used to generate a final target detection result.
[0009] Furthermore, the step of inputting the labeled sample image into the YOLO target detection network for training to obtain a trained target detection model includes: Set the hyperparameters of the YOLO target detection network; wherein the hyperparameters include BatchSize and Epoch; Divide the labeled sample images into a training set, a validation set, and a test set in a ratio of 7:2:1; The training set and the validation set are sent to the YOLO target detection network for training to obtain a plurality of converged target detection models; The target detection model is tested using the test set, and the target detection model with the best recognition effect is selected so that the target detection model can recognize the outer contour of the oil cup and the liquid level line area.
[0010] Furthermore, the step of using a camera to collect a large field of view image containing the constant position oil cup with limit glass to be identified at a preset inspection point includes: The inspection robot carrying the camera arrives at the inspection point according to a preset inspection route; The inspection robot adjusts the camera to the azimuth angle and pitch angle preset at the inspection point, and captures the constant-position oil cup with limit glass to obtain a large-field image containing the constant-position oil cup with limit glass to be identified.
[0011] Furthermore, the step of using the trained target detection model to recognize the small field of view image to obtain the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass includes: The small field of view image is recognized by using the trained target detection model to obtain multiple sets of second outer contour position information and liquid level line position information of the constant position oil cup with limit glass; According to the inherent positional relationship between the lowest liquid level, the highest liquid level and the liquid level line in the constant level oil cup with limit glass, as well as the confidence of the outer contour of the oil cup and the liquid level line, invalid recognition results of the outer contour position and the liquid level line position are arbitrated and deleted to obtain a set of accurate second outer contour position information and the liquid level line position information.
[0012] Further, the step of obtaining the oil cup liquid level status information according to the instrument prior information, the second outer contour position information and the liquid level line position information includes: According to the second outer contour position information of the constant level oil cup with limit glass and the liquid level line position information, combined with the instrument prior information of the inspection point, calculate the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level; According to the second outer contour position information of the constant level oil cup with limit glass and the position information of the liquid level line, combined with the prior information of the instrument at the inspection point, the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level is calculated; The oil cup liquid level status information is obtained according to the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level, and the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level.
[0013] Furthermore, the minimum liquid level is the liquid level position when the oil volume in the constant level oil cup with a limit glass is the lowest, the maximum liquid level is the maximum oil volume contained in the constant level oil cup with a limit glass, and the instrument prior information includes information such as the minimum liquid level, the maximum liquid level, the relative position between the limit and the outer contour.
[0014] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects: In the embodiments of the present disclosure, through the above-mentioned image-based method for monitoring the liquid level status information of the constant position oil cup with limit glass, on the one hand, the image of the constant position oil cup with limit glass is annotated; the annotated image data is sent to the YOLO target detection network for training to obtain a target detection model for identifying the outer contour area of the oil cup and the liquid level line area; the inspection robot gimbal collects a large field of view image of the constant position oil cup with limit glass; the inspection robot gimbal is adjusted to obtain a small field of view image of the constant position oil cup with limit glass through the result of detecting the large field of view image of the constant position oil cup with limit glass by the target detection model; the small field of view image of the constant position oil cup with limit glass is detected through the target detection model to obtain the outer contour of the oil cup and the liquid level line position; according to the prior information of the constant position oil cup with limit glass detected by the target detection model, the digital information result of the liquid level status of the constant position oil cup with limit glass is calculated. On the other hand, it effectively overcomes the problem of low efficiency of manual monitoring, realizes automatic identification, and improves the efficiency and accuracy of liquid level status information monitoring; it adopts advanced deep learning algorithms to effectively deal with the interference of external lighting environment, and significantly enhances the versatility and robustness of the algorithm; it is suitable for various types of constant-level oil cups with limit glass, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.
[0016] Figure 1 A step diagram showing a method for monitoring the liquid level status information of a constant-position oil cup with a limit glass based on an image in an exemplary embodiment of the present disclosure; Figure 2 An example image is shown in which a large field of view image containing a constant-position oil cup with a limit glass to be identified is captured by an inspection robot in an exemplary embodiment of the present disclosure; Figure 3 An example image is shown in which an inspection robot in an exemplary embodiment of the present disclosure captures a small field of view image containing a constant-position oil cup with a limit glass to be identified. DETAILED DESCRIPTION
[0017] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
[0018] In addition, the accompanying drawings are only schematic illustrations of the embodiments of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated descriptions will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities.
[0019] This example embodiment provides an image-based method for monitoring the liquid level status information of a constant-position oil cup with a limit glass. Figure 1 As shown in , the image-based method for monitoring the liquid level status information of the constant-position oil cup with a limit glass may include: steps S101 to S107.
[0020] Step S101: acquiring a sample image, and marking the outer contour of the oil cup and the liquid level line area of the constant-level oil cup with a limiting glass in the sample image; Step S102: inputting the labeled sample image into the YOLO target detection network for training to obtain a trained target detection model; Step S103: using a camera to collect a large field of view image including the constant position oil cup with limit glass to be identified at a preset inspection point; Step S104: using the trained target detection model to identify the large field of view image, and obtaining first outer contour position information of the constant position oil cup with limit glass; Step S105: adjusting the angle of the camera according to the first outer contour position information and re-collecting a small field of view image including the constant position oil cup with limit glass to be identified; Step S106: using the trained target detection model to identify the small field of view image, and obtaining the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass; Step S107: Obtaining the oil cup liquid level status information according to the instrument prior information, the second outer contour position information and the liquid level line position information.
[0021] Through the above-mentioned image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass, on the one hand, the image of the constant-position oil cup with limit glass is annotated; the annotated image data is sent to the YOLO target detection network for training to obtain a target detection model for identifying the outer contour area and the liquid level line area of the oil cup; the inspection robot gimbal collects a large-field-of-view image of the constant-position oil cup with limit glass; the inspection robot gimbal is adjusted to obtain a small-field-of-view image of the constant-position oil cup with limit glass through the result of detecting the large-field-of-view image of the constant-position oil cup with limit glass by the target detection model; the small-field-of-view image of the constant-position oil cup with limit glass is detected through the target detection model to obtain the outer contour of the oil cup and the position of the liquid level line; according to the prior information of the constant-position oil cup with limit glass detected by the target detection model, the digital information result of the liquid level status of the constant-position oil cup with limit glass is calculated. On the other hand, it effectively overcomes the problem of low efficiency of manual monitoring, realizes automatic identification, and improves the efficiency and accuracy of liquid level status information monitoring; it adopts advanced deep learning algorithms to effectively deal with the interference of external lighting environment, and significantly enhances the versatility and robustness of the algorithm; it is suitable for various types of constant-level oil cups with limit glass, and has broad application prospects.
[0022] Next, we will refer to Figures 1 to 3 The various steps of the above-mentioned image-based method for monitoring the liquid level status information of the constant-level oil cup with limit glass in this example implementation are described in more detail.
[0023] In step S101, a sample image is acquired, and the outer contour of the oil cup and the liquid level line area of the constant-level oil cup with limiting glass in the sample image are marked.
[0024] Specifically, the pre-collected images of the constant-position oil cup with limit glass are subjected to data cleaning and data augmentation, wherein data cleaning mainly includes removing duplicate images, missing images, useless images, etc., and data augmentation mainly includes image rotation, image mirroring, etc.
[0025] Use the image annotation tool software to annotate the outer contour of the oil cup and the liquid level line area in the image of the constant-position oil cup with limit glass; the annotation tool software supports multiple annotation formats and is suitable for annotating image datasets for deep learning model training.
[0026] In step S102, the labeled sample image is input into the YOLO target detection network for training to obtain a trained target detection model.
[0027] Specifically, the target detection network based on deep learning has powerful feature extraction capabilities, among which the YOLO (YouOnly Look Once) target detection network has the characteristics of open source, high recognition accuracy, and fast speed. For this reason, this method divides the shuffled annotated image dataset into training set, validation set, and test set in a ratio of 7:2:1, and uses the deep learning network YOLO for model training to obtain a model for identifying the outer contour of the oil cup and the liquid level line area.
[0028] More specifically, a YOLO target detection network is first constructed, wherein the YOLO target detection network specifically includes an input end, a Backbone, a Neck, and a prediction output end. The input end of the YOLO target detection network is subjected to image preprocessing for model input. The image preprocessing performed on the input end of the YOLO target detection network includes image size adjustment, RGB value normalization of image pixels, and image data enhancement. Backbone uses Cross Stage Partial Darknet 53, Neck includes Path Aggregation Network and Feature Pyramid Network, and the prediction output end includes a detection head, which includes a convolutional layer, a candidate box generation layer, a prediction layer, and an activation function for generating a final target detection result.
[0029] The labeled data set is sent to the YOLO deep learning network (i.e., YOLO target detection network) for training to obtain a model for identifying the outer contour of the oil cup and the liquid level line area. First, the network model hyperparameters are set, mainly including BatchSize, Epoch, etc., and then the training set and validation set are used to train until the model converges. Finally, the YOLO weight with the best recognition effect is selected through the test set test.
[0030] In step S103, a camera is used to collect a large-field image including the constant-position oil cup with limit glass to be identified at a preset inspection point.
[0031] Specifically, the robot arrives at the inspection point according to the inspection route, and adjusts the camera to the pre-set azimuth and pitch angle of the inspection point to capture images of the constant-position oil cup with limit glass at the inspection point, and collects a large-field-of-view image of the constant-position oil cup with limit glass to be identified.
[0032] In step S104 and step S105, the large field of view image is identified using the trained target detection model to obtain the first outer contour position information of the constant position oil cup with limit glass; the angle of the camera is adjusted according to the first outer contour position information and the small field of view image containing the constant position oil cup with limit glass to be identified is recaptured.
[0033] Specifically, the trained target detection model is used to identify the large field of view image with the constant position oil cup of the limit glass, and the position information of the constant position outer contour of the limit glass is obtained, including the coordinates of its center point in the image ( , Based on this coordinate and the coordinate of the center point of the large field of view image ( , ), combined with the camera's zoom factor Z, image resolution W and H, and maximum horizontal field of view And the maximum vertical field of view , adjust the azimuth and pitch angles of the camera using the following formula:
[0034]
[0035] in, and Respectively represent the current azimuth and pitch angle of the camera. and pitch angle , reposition the direction of the camera so that the outer contour of the constant oil cup with limit glass appears in the center of the image. Based on the camera, a small field of view image containing the constant oil cup with limit glass to be identified is collected.
[0036] In step S106, the trained target detection model is used to identify the small field of view image to obtain the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass.
[0037] Specifically, the trained YOLO model is used to identify the small field of view image to obtain the outer contour and liquid level line position information of the constant-level oil cup with limit glass; according to the inherent position relationship between the lowest liquid level, the highest liquid level and the liquid level line in the constant-level oil cup with limit glass, as well as the confidence of the outer contour and the liquid level line of the oil cup, the invalid recognition results of the outer contour and the liquid level line of the oil cup are arbitrated and deleted to obtain a set of accurate position information of the outer contour and the liquid level line of the oil cup.
[0038] More specifically, according to the small field image of the constant position oil cup with limit glass, the recognition software module uses the trained model for recognizing the outer contour of the oil cup and the liquid level line area to recognize the small field image of the constant position oil cup with limit glass, and obtains the position information of the outer contour of the oil cup and the liquid level line; According to the inherent positional relationship between the lowest liquid level, the highest liquid level and the liquid level line in the constant-level oil cup with limit glass, as well as the confidence of the outer contour of the oil cup and the liquid level line, invalid recognition results of the outer contour of the oil cup and the liquid level line are arbitrated and deleted to obtain a set of accurate position information of the outer contour of the oil cup and the liquid level line.
[0039] In step S107, the oil cup liquid level status information is obtained according to the instrument prior information, the second outer contour position information and the liquid level line position information, and the formula is as follows:
[0040] in, and are the heights of the highest liquid level and the lowest liquid level in the image, respectively, is the height of the liquid level line in the image.
[0041] According to the outer contour of the oil cup with limit glass constant level oil cup and the position information of the liquid level line, combined with the prior information of the instrument at the inspection point, the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level is calculated; According to the outer contour of the oil cup with limit glass constant level oil cup and the position information of the liquid level line, combined with the prior information of the instrument at the inspection point, the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level is calculated; According to the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level, and the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level, the calculation result outputs the final digital information result of the oil cup liquid level status (i.e., the oil cup liquid level status information).
[0042] It can be understood that the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level, and the size relationship between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level are used to jointly represent the liquid level status information. Among them, the lowest liquid level refers to the liquid level position when the oil cup has the lowest oil volume, and the highest liquid level refers to the maximum oil volume that the oil cup can hold. The instrument prior information is based on manual calibration.
[0043] In a specific embodiment, after the target detection model is trained, the robot arrives at the inspection point according to the inspection route, and the robot adjusts the camera to the pre-set azimuth and pitch angle of the inspection point to capture images of the constant position oil cup with limit glass at the inspection point, and collects a large field of view image containing the constant position oil cup with limit glass to be identified, such as Figure 2 shown.
[0044] The trained YOLO model is used to identify the large field of view image of the constant position oil cup with limit glass, and the position information of the outer contour of the constant position oil cup with limit glass is obtained. The azimuth and pitch angle of the camera are adjusted according to the outer contour position information so that the outer contour of the constant position oil cup with limit glass is in the center of the image; based on the small field of view image of the constant position oil cup with limit glass to be identified collected by the camera, such as Figure 3 shown.
[0045] The small field of view image is recognized using the trained YOLO model to obtain the outer contour and liquid level position information of the constant-position oil cup with limit glass; Combined with the prior information of the instrument at the inspection point, the outer contour of the oil cup, and the position information of the liquid level line, the digital information of the oil cup level status is finally obtained by calculating the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level, and the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level.
[0046] Through the above-mentioned image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass, on the one hand, the image of the constant-position oil cup with limit glass is annotated; the annotated image data is sent to the YOLO target detection network for training to obtain a target detection model for identifying the outer contour area and the liquid level line area of the oil cup; the inspection robot gimbal collects a large-field-of-view image of the constant-position oil cup with limit glass; the inspection robot gimbal is adjusted to obtain a small-field-of-view image of the constant-position oil cup with limit glass through the result of detecting the large-field-of-view image of the constant-position oil cup with limit glass by the target detection model; the small-field-of-view image of the constant-position oil cup with limit glass is detected through the target detection model to obtain the outer contour of the oil cup and the position of the liquid level line; according to the prior information of the constant-position oil cup with limit glass detected by the target detection model, the digital information result of the liquid level status of the constant-position oil cup with limit glass is calculated. On the other hand, it effectively overcomes the problem of low efficiency of manual monitoring, realizes automatic identification, and improves the efficiency and accuracy of liquid level status information monitoring; it adopts advanced deep learning algorithms to effectively deal with the interference of external lighting environment, and significantly enhances the versatility and robustness of the algorithm; it is suitable for various types of constant-level oil cups with limit glass, and has broad application prospects.
[0047] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.
[0048] In the embodiments of the present disclosure, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present disclosure can be understood according to specific circumstances.
[0049] In the embodiments of the present disclosure, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include the first and second features being in direct contact, or may include the first and second features not being in direct contact but being in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes the first feature being directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes the first feature being directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.
[0050] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification.
[0051] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.
Claims
1. An image-based method for monitoring the liquid level status information of a constant-position oil cup with a limit glass, characterized in that: The method includes: Acquire a sample image, and mark the outer contour of the oil cup and the liquid level line area of the constant-level oil cup with limiting glass in the sample image; Input the labeled sample image into the YOLO target detection network for training to obtain a trained target detection model; Using a camera to collect a large field of view image containing the constant position oil cup with limit glass to be identified at a preset inspection point; Using the trained target detection model to identify the large field of view image, obtaining first outer contour position information of the constant position oil cup with limit glass; Adjust the angle of the camera according to the first outer contour position information and re-capture a small field of view image including the constant position oil cup with limit glass to be identified; Using the trained target detection model to identify the small field of view image, obtain the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass; The oil cup liquid level status information is obtained according to the instrument priori information, the second outer contour position information and the liquid level line position information.
2. According to claim 1, the image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass is characterized in that: The step of acquiring a sample image and marking the outer contour of the oil cup and the liquid level line area of the constant-position oil cup with a limiting glass in the sample image includes: Acquire the sample image, and perform data cleaning on the sample image to remove duplicate images, missing images and useless images; The outer contour of the oil cup and the liquid level line area of the sample image after cleaning are marked.
3. According to claim 1, the image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass is characterized in that: The method further includes: Constructing a YOLO target detection network; wherein the YOLO target detection network includes an input end, a backbone, a neck, and a prediction output end; Performing image preprocessing on the input end of the YOLO target detection network for model input, wherein the image preprocessing includes image size adjustment, RGB value normalization of image pixels, and image data enhancement; The Backbone is Cross Stage Partial Darknet 53; The Neck includes a Path Aggregation Network and a Feature Pyramid Network. The prediction output end includes a detection head, which includes a convolution layer, a candidate box generation layer, a prediction layer and an activation function. The prediction output end is used to generate a final target detection result.
4. According to claim 3, the image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass is characterized in that: The step of inputting the labeled sample image into the YOLO target detection network for training to obtain a trained target detection model includes: Set the hyperparameters of the YOLO target detection network; wherein the hyperparameters include BatchSize and Epoch; Divide the labeled sample images into a training set, a validation set, and a test set in a ratio of 7:2:1; The training set and the validation set are sent to the YOLO target detection network for training to obtain a plurality of converged target detection models; The target detection model is tested using the test set, and the target detection model with the best recognition effect is selected so that the target detection model can recognize the outer contour of the oil cup and the liquid level line area.
5. According to claim 1, the image-based method for monitoring the liquid level status information of the constant-position oil cup with limit glass is characterized in that: The step of using a camera to collect a large field of view image containing a constant position oil cup with a limit glass to be identified at a preset inspection point includes: The inspection robot carrying the camera arrives at the inspection point according to a preset inspection route; The inspection robot adjusts the camera to the azimuth angle and pitch angle preset at the inspection point, and captures the constant-position oil cup with limit glass to obtain a large-field image containing the constant-position oil cup with limit glass to be identified.
6. The image-based method for monitoring the liquid level status information of a constant-position oil cup with a limit glass according to claim 1 is characterized in that: The step of using the trained target detection model to identify the small field of view image to obtain the second outer contour position information and the liquid level line position information of the constant position oil cup with limit glass includes: The small field of view image is recognized by using the trained target detection model to obtain multiple sets of second outer contour position information and liquid level line position information of the constant position oil cup with limit glass; According to the inherent positional relationship between the lowest liquid level, the highest liquid level and the liquid level line in the constant level oil cup with limit glass, as well as the confidence of the outer contour of the oil cup and the liquid level line, invalid recognition results of the outer contour position and the liquid level line position are arbitrated and deleted to obtain a set of accurate second outer contour position information and the liquid level line position information.
7. The image-based method for monitoring the liquid level status information of a constant-position oil cup with a limit glass according to claim 1 is characterized in that: The step of obtaining the oil cup liquid level status information according to the instrument prior information, the second outer contour position information and the liquid level line position information includes: According to the second outer contour position information of the constant level oil cup with limit glass and the liquid level line position information, combined with the instrument prior information of the inspection point, calculate the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level; According to the second outer contour position information of the constant level oil cup with limit glass and the position information of the liquid level line, combined with the prior information of the instrument at the inspection point, the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level is calculated; The oil cup liquid level status information is obtained according to the ratio of the distance from the liquid level line to the lowest liquid level to the distance from the highest liquid level to the lowest liquid level, and the difference between the distance from the liquid level line to the lowest liquid level and the distance from the limit to the lowest liquid level.
8. The image-based method for monitoring the liquid level status information of a constant-position oil cup with a limit glass according to claim 7 is characterized in that: The minimum liquid level is the liquid level position when the oil volume in the constant level oil cup with limit glass is the lowest, and the maximum liquid level is the maximum oil volume contained in the constant level oil cup with limit glass. The instrument prior information includes information such as the minimum liquid level, the maximum liquid level, the relative position between the limit and the outer contour.