Early warning method based on visual detection of liquid leakage of metering pump

Abstract

The invention discloses an early warning method based on visual detection of liquid leakage of a metering pump, and relates to the technical field of liquid leakage detection of the metering pump, and the method comprises the steps: S1, collecting an image of a working area of the metering pump in real time through an image collection module; s2, preprocessing the acquired image through an image processing module; S3, outputting by using a neural network to obtain an original output vector z; s4, calculating a classification loss value according to a cross entropy loss function; s5, according to a preset loss threshold value, triggering an early warning signal of a corresponding grade; s6, the detection result is displayed in real time through intelligent analysis software, and liquid leakage event information is recorded. According to the invention, leakage grading early warning is carried out on the metering pump, so that maintenance, rescue and other resources can be reasonably allocated, low-grade leakage can be handled by arranging less manpower and material resources, high-grade leakage needs to be handled by centralizing a large amount of resources, and resource waste or insufficiency is avoided.

Description

Technical Field

[0001] The invention relates to the technical field of metering pump leakage detection, in particular to an early warning method based on visual detection of metering pump leakage. Background Art

[0002] In industrial production, metering pumps are key equipment for transporting liquids, and their stable operating status is crucial to production safety and efficiency. Leakage is a common fault in metering pumps. If not detected and addressed promptly, it can lead to material waste, equipment damage, and even safety accidents. Traditional leakage detection methods typically rely on manual inspections, which suffer from low detection efficiency, strong subjectivity, and poor real-time performance. These methods are unable to meet the requirements of modern industrial automation production for equipment safety monitoring. Therefore, visual detectors are now commonly used to detect leakage in metering pumps.

[0003] Since different degrees of leakage pose different risks to personnel safety, the environment and equipment, for example, mild leakage may only require enhanced monitoring, while severe leakage requires immediate evacuation of personnel and activation of emergency plans. However, the existing metering pump leakage warning method based on visual sensors is single and cannot provide graded alarms based on the leakage situation of the metering pump, which can easily lead to waste or insufficient resources.

[0004] Based on this, an early warning method based on visual detection of metering pump leakage is now provided, which can eliminate the disadvantages of existing devices. Summary of the Invention

[0005] The purpose of the present invention is to provide an early warning method based on visual detection of metering pump leakage, so as to solve the problem of the singleness of the existing metering pump leakage early warning method based on visual sensors in the background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An early warning method based on visual detection of metering pump leakage includes the following steps:

[0008] S1: Real-time acquisition of images of the working area of ​​the metering pump through the image acquisition module;

[0009] S2: Preprocessing the collected image through the image processing module to extract the shape and area features of the liquid area;

[0010] S3: Using the neural network output, the original output vector z is obtained. Vector z is the input vector of the improved Softmax function. Assume that the input vector is z = [z1, z2, z3], where z1, z2, z3 represent the area of ​​the liquid in the image. After the improved Softmax function is processed, the output vector y = [y1, y2, y3] is obtained, where y1, y2, y3 are the probability distributions of different images in different areas. The theoretical calculation formula is

[0011] S4: Calculate the classification loss value according to the cross entropy loss function. The formula is: Loss = -log(p k ), where k is the index of the true category;

[0012] S5: According to the preset loss threshold, trigger the corresponding level of warning signal, and transmit the signal to the alarm notification module, triggering response measures such as sound and light alarm, SMS notification, etc.

[0013] S6: Intelligent analysis software is used to display the test results in real time, record leakage event information, and provide a user interaction interface.

[0014] On the basis of the above technical solutions, the present invention also provides the following optional technical solutions:

[0015] In one optional solution, the corresponding relationship between different levels of warning signals and loss values ​​is as follows:

[0016] When k=1, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more general warning it is.

[0017] When k=2, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the medium warning.

[0018] When k=3, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more urgent the warning.

[0019] In an optional solution: in step S1, the image acquisition module is composed of an industrial camera, which is used to obtain images of the working area of ​​the metering pump in real time.

[0020] In an optional solution: the image processing module in step S2 includes a preprocessing unit and a classification warning algorithm unit, which are used to extract image features and perform improved Softmax function and cross entropy loss calculation.

[0021] In an optional solution: the alarm notification module in step S5 includes an audible and visual alarm and a text message notification interface, which is used to trigger a corresponding response according to the warning level.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. The present invention uses advanced visual detection technology to monitor the working status of the metering pump in real time, especially to accurately identify and alarm possible leakage. It combines industrial cameras, image processing algorithms and intelligent analysis software to efficiently and accurately detect leakage, providing reliable safety protection for industrial production.

[0024] 2. The present invention can reasonably allocate resources such as maintenance and emergency rescue by providing graded early warning of leakage of metering pumps. For low-level leakage, less manpower and material resources can be arranged for processing, while for high-level leakage, a large amount of resources need to be concentrated for emergency disposal to avoid waste or shortage of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flowchart of the early warning method steps of the present invention. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0027] In one embodiment, Figure 1 As shown, an early warning method based on visual detection of metering pump leakage includes the following steps: S1: collecting images of the working area of ​​the metering pump in real time through an image acquisition module;

[0028] S2: Preprocessing the collected image through the image processing module to extract the shape and area features of the liquid area;

[0029] S3: Using the neural network output, the original output vector z is obtained. Vector z is the input vector of the improved Softmax function. Assume that the input vector is z = [z1, z2, z3], where z1, z2, z3 represent the area of ​​the liquid in the image. After the improved Softmax function is processed, the output vector y = [y1, y2, y3] is obtained, where y1, y2, y3 are the probability distributions of different images in different areas. The theoretical calculation formula is

[0030] S4: Calculate the classification loss value according to the cross entropy loss function. The formula is: Loss = -log(p k ), where k is the index of the true category;

[0031] S5: According to the preset loss threshold, trigger the corresponding level of warning signal, and transmit the signal to the alarm notification module, triggering response measures such as sound and light alarm, SMS notification, etc.

[0032] S6: Intelligent analysis software is used to display the test results in real time, record leakage event information, and provide a user interaction interface.

[0033] In one embodiment, the corresponding relationship between different levels of warning signals and loss values ​​is as follows:

[0034] When k=1, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more general warning it is.

[0035] When k=2, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the medium warning.

[0036] When k=3, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more urgent the warning.

[0037] In one embodiment, the image acquisition module in step S1 is composed of an industrial camera, which is used to acquire images of the working area of ​​the metering pump in real time.

[0038] In one embodiment, the image processing module in step S2 includes a preprocessing unit and a classification warning algorithm unit, which are used to extract image features and perform improved Softmax function and cross entropy loss calculation.

[0039] In one embodiment, the alarm notification module in step S5 includes an audible and visual alarm, a text message notification interface, etc., which is used to trigger a corresponding response according to the warning level.

[0040] The above embodiment discloses an early warning method based on visual detection of metering pump leakage, and its specific working principle and process are as follows:

[0041] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An early warning method based on visual detection of metering pump leakage, characterized in that: The following steps are involved: S1: Real-time acquisition of images of the working area of ​​the metering pump through the image acquisition module; S2: Preprocessing the collected image through the image processing module to extract the shape and area features of the liquid area; S3: Using the neural network output, the original output vector z is obtained. Vector z is the input vector of the improved Softmax function. Assume that the input vector is z = [z1, z2, z3], where z1, z2, z3 represent the area of ​​the liquid in the image. After the improved Softmax function is processed, the output vector y = [y1, y2, y3] is obtained, where y1, y2, y3 are the probability distributions of different images in different areas. The theoretical calculation formula is S4: Calculate the classification loss value according to the cross entropy loss function. The formula is: Loss = -log(p k ), where k is the index of the true category; S5: According to the preset loss threshold, trigger the corresponding level of warning signal, and transmit the signal to the alarm notification module, triggering response measures such as sound and light alarm, SMS notification, etc. S6: Intelligent analysis software is used to display the test results in real time, record leakage event information, and provide a user interaction interface.

2. The early warning method based on visual detection of metering pump leakage according to claim 1 is characterized in that: The corresponding relationship between different levels of warning signals and loss values ​​is: When k=1, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more general warning it is. When k=2, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the medium warning. When k=3, the vector z is converted into a probability distribution p by improving the Softmax function, and then the loss is calculated based on the cross entropy. The smaller the value, the more urgent the warning.

3. The early warning method based on visual detection of metering pump leakage according to claim 1 is characterized in that: In step S1, the image acquisition module is composed of an industrial camera and is used to obtain images of the working area of ​​the metering pump in real time.

4. The early warning method based on visual detection of metering pump leakage according to claim 1 is characterized in that: The image processing module in step S2 includes a preprocessing unit and a classification warning algorithm unit, which are used to extract image features and perform improved Softmax function and cross entropy loss calculation.

5. The early warning method based on visual detection of metering pump leakage according to claim 1 is characterized in that: The alarm notification module in step S5 includes an audible and visual alarm and a text message notification interface, which is used to trigger a corresponding response according to the warning level.

Images