Meteorological detection method based on image recognition and cold mirror dew point meter for detection

Abstract

The present application relates to a kind of meteorological detection method based on image recognition and for detecting cold mirror dew point instrument, mirror surface is cleaned, mirror surface image is collected, calculates dew condensation brightness determination value, determines easy dew condensation area, according to real-time measurement data dynamic PID control, obtains the temperature when dew point condenses and the temperature when dew point disappears, finally two temperatures are fused to obtain accurate dew point temperature.The present application is high in degree of automation, and detection precision is good.

Description

Technical Field

[0001] This invention relates to a meteorological detection method based on image recognition and a cold mirror dew point meter for detection, belonging to the field of meteorological detection technology. Background Technology

[0002] With the development of industry and science and technology, the monitoring of gas and water vapor content has become increasingly important. Many industrial and scientific fields, such as meteorology, chemical industry, and electronics manufacturing, require accurate measurement of the water vapor content in the air. Dew point is a crucial indicator for measuring the water vapor content in the air. Dew point temperature refers to the temperature at which water vapor begins to condense into dew after the air becomes saturated with water under constant pressure and the temperature drops to a certain value. Therefore, dew point temperature can be used to reflect the water vapor content in the air.

[0003] A cold mirror dew point meter is a commonly used instrument for measuring dew point temperature. Its basic principle is to cool air to a temperature below the dew point, causing water vapor to condense into water droplets on a condenser. A camera then captures images of these droplets on the cold mirror, and image processing algorithms analyze parameters such as the droplet shape and number to calculate the dew point temperature. Traditional cold mirror dew point meters require manual operation, demanding a high level of skill from the operator and are susceptible to human error. To improve measurement accuracy and automation, many modern cold mirror dew point meters employ automatic control systems that use image recognition technology to automatically identify water droplets and perform automatic measurements; however, the image recognition technology used in these systems cannot guarantee the accuracy and reliability of the dew point measurement.

[0004] Therefore, it is urgent to focus research on cold mirror dew point meters, which can improve both automation and detection accuracy. Summary of the Invention

[0005] This invention provides a meteorological detection method based on image recognition and a cold mirror dew point meter for detection, which has a high degree of automation and excellent detection accuracy.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A meteorological detection method based on image recognition specifically includes the following steps:

[0008] Step S1: Start the cold mirror dew point meter. High-pressure gas is sprayed onto the cold mirror surface for cleaning. After cleaning is completed, the refrigeration unit applies a reverse voltage to the bottom of the cold mirror to heat it until the cold mirror surface is dry.

[0009] Step S2: Simultaneously activate the LED and camera, input the gas to be tested into the gas chamber, the camera acquires real-time mirror images of the cold mirror, and sends the acquired mirror images to the main control room; select the detection image from several mirror images;

[0010] Step S3: Take multiple measurements and, based on the selected detection images, calculate the variance σ1 of the brightness change caused by detection error when the cold mirror surface is free of condensation;

[0011] Step S4: Cool the air chamber for the first time, start the dynamic PID controller, and adjust the parameters of the dynamic PID controller according to the deviation between the real-time acquired detection image brightness L and the brightness judgment target value mσ1.

[0012] Step S5: When the temperature inside the air chamber decreases until the brightness of the detected image reaches the target value of the brightness judgment value, i.e., L=mσ1, the area prone to condensation is identified, and the cooling of the air chamber is stopped;

[0013] Step S6: Mark all pixels in the condensation-prone areas of the detection image as 1 and the rest as 0, and call it the template image. Perform a bitwise AND operation between the detection image and the template image to obtain a detection image that retains only the condensation-prone areas, and call it the ROI region detection image.

[0014] Step S7: Measure and calculate the brightness change caused by detection error in the ROI region detection image when there is no condensation. Calculate its variance σ2 multiple times and redetermine the target value mσ2 of the brightness judgment value as the brightness threshold value for judging condensation.

[0015] Step S8: Cool the air chamber a second time, restart the dynamic PID controller, adjust the parameters of the dynamic PID controller according to the deviation between the real-time detected image brightness L and the brightness judgment target value mσ2, calculate the brightness F of the detected image of the ROI area, and stop cooling immediately when the system cools down to the brightness judgment target value, i.e. F=mσ2, and record the temperature at this time as the dew point temperature T1.

[0016] Step S9: The temperature of the air chamber gradually increases, and the dew point gradually disappears. When the temperature inside the air chamber rises to F = σ2, record the temperature at this time as the dew point temperature T2.

[0017] Step S10: Calculate the dew point temperature by combining the dew point temperature T1 and the dew point temperature T2 with the set weights, and the detection ends;

[0018] As a further preferred embodiment of the present invention, in step S2, the 10 frames of images when the cold mirror surface is free of condensation are accumulated and averaged to obtain the background image, and the subsequent 10 frames of images are accumulated and averaged and the background image is subtracted to obtain the detection image.

[0019] As a further preferred embodiment of the present invention, in step S3, multiple measurements are taken, and the variance σ1 of the brightness change caused by detection error when there is no condensation is calculated. The calculation formula is as follows:

[0020]

[0021] Where L is the brightness of the detected image acquired in real time, L=R*0.299+G*0.587+B*0.114, R, G, B are the values ​​of the red, green and blue color channels in the RGB color space, respectively; The value is the average brightness variation; n is the number of measurements.

[0022] mσ1 is taken as the target value for the brightness determination of condensation in the cold mirror dew point meter. Based on the experiment, m = 6 is determined, and an allowable fluctuation range of the brightness L of the detection image is set according to the maximum and minimum values ​​of the measured data.

[0023] As a further preferred embodiment of the present invention, in step S4, when the deviation is large, the proportional coefficient is increased, the integral time constant is decreased, and the derivative time constant is increased to improve the response speed and stability of temperature control; when the deviation is small, the opposite is true.

[0024] As a further preferred embodiment of the present invention, in step S5, when the temperature inside the air chamber decreases until L = 6σ1, the current detection image is subjected to median filtering, a 2x2 filter kernel is selected to remove interference from two or fewer pixels, and the scattered dew points are identified as areas prone to condensation. The cold mirror dew point meter immediately stops cooling.

[0025] As a further preferred embodiment of the present invention, in step S10, the dew point temperatures T1 and T2 read out respectively are fused to calculate the measured dew point temperature value. The calculation formula is as follows:

[0026] T = kT1 + (1-k)T2

[0027] Where k is the weighting coefficient, and k = 0.7 was determined based on experiments;

[0028] The cold mirror dew point meter used in the image recognition-based meteorological detection method includes a box for setting up the main control room, a refrigeration stack is set up at the center of the top of the box, a gas chamber is located at the top of the refrigeration stack, and a cold mirror is embedded in the gas chamber relative to the bottom of the refrigeration stack, and a platinum resistance thermometer is embedded at the center of the cold mirror.

[0029] A camera is mounted on top of the air chamber, with its lens facing the center of the cold mirror to capture images of the mirror surface.

[0030] An air inlet and an air outlet are respectively set on both sides of the air chamber; LEDs are also installed on the side wall where the air outlet is located, and the light emitted by the LEDs is directed toward the cold mirror surface at a set incident angle.

[0031] A heat dissipation component combining air cooling and water cooling is installed at the bottom of the enclosure;

[0032] A touchscreen is installed on one side of the enclosure;

[0033] The main control unit is connected to the refrigeration unit, camera, touchscreen, and heat dissipation components.

[0034] As a further preferred embodiment of the present invention, the incident angle is 50°.

[0035] By employing the above technical solutions, the present invention has the following beneficial effects compared to the prior art:

[0036] 1. The meteorological detection method based on image recognition provided by the present invention first locates the area prone to condensation, then performs brightness detection on the area prone to condensation, and integrates the temperature at the time of condensation and the temperature at the time of dew point disappearance, which greatly improves the accuracy and reliability of dew point detection.

[0037] 2. The cold mirror dew point meter of the matching detection method provided by the present invention captures images of the mirror surface with a camera, and determines whether condensation has formed on the mirror surface based on the changes in pixel brightness of the captured image, thus laying the foundation for the accuracy of dew point detection.

[0038] 3. The image recognition-based meteorological detection method provided by this invention can effectively prevent equipment from falsely detecting dew points, improve the reliability and stability of production equipment, reduce the workload of manual inspection and maintenance, lower production costs and improve efficiency. Attached Figure Description

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0040] Figure 1 This is a schematic diagram of the overall appearance of the cold mirror dew point meter provided by the present invention;

[0041] Figure 2 This is an overall side view of the cold mirror dew point meter provided by the present invention;

[0042] Figure 3 This is a side cross-sectional view of the cold mirror dew point meter provided by the present invention;

[0043] Figure 4 This is a flowchart of the meteorological detection method based on image recognition provided by the present invention.

[0044] In the diagram: 1 is the camera, 2 is the air chamber, 3 is the air outlet, 4 is the light-emitting diode, 5 is the touch screen, 6 is the refrigeration stack, 7 is the main control room, 8 is the heat dissipation component, 9 is the cold mirror, 10 is the platinum resistance thermometer, and 11 is the air inlet. Detailed Implementation

[0045] The present invention will now be described in further detail with reference to the accompanying drawings. In the description of this application, it should be understood that the terms "left side," "right side," "upper part," "lower part," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. "First," "second," etc., do not indicate the importance of the components, and therefore should not be construed as a limitation of the present invention. The specific dimensions used in this embodiment are only for illustrating the technical solution and do not limit the scope of protection of the present invention.

[0046] As described in the background section, traditional cold mirror dew point meters require manual operation, demanding high skill levels from the operator and are susceptible to human error. Most importantly, the image recognition technology upon which traditional cold mirror dew point meters rely has low accuracy in automatically identifying water droplets, resulting in compromised measurement accuracy and reliability.

[0047] Therefore, this application first refines and updates the measurement method of traditional cold mirror dew point meters. It mainly involves first locating areas prone to condensation, then measuring the brightness of these areas, and combining the temperature at which condensation occurs with the temperature at which the dew point disappears, greatly improving the accuracy and reliability of dew point detection. Figure 4 As shown, the specific steps include:

[0048] Step S1: Start the cold mirror dew point meter. High-pressure gas is sprayed onto the surface of the cold mirror 9 for cleaning. After cleaning is completed, the refrigeration stack 6 applies a reverse voltage to the bottom of the cold mirror to heat it until the surface of the cold mirror is dry.

[0049] Step S2: Simultaneously activate LED 4 and camera 1, input the gas to be tested into gas chamber 2, and the camera acquires real-time mirror images of the cold mirror. Send the acquired mirror images to the main control room 7. The 10 frames of images when the cold mirror surface is free of condensation are accumulated and averaged to obtain the background image. The subsequent 10 frames of images are accumulated and averaged, and the background image is subtracted to obtain the detection image.

[0050] Step S3: After multiple measurements, based on the selected detection image, calculate the variance σ1 of the brightness change caused by detection error when the cold mirror surface is free of condensation. The calculation formula is as follows:

[0051]

[0052] Where L is the brightness of the detected image acquired in real time, L=R*0.299+G*0.587+B*0.114, R, G, B are the values ​​of the red, green and blue color channels in the RGB color space, respectively; The value is the average brightness variation; n is the number of measurements.

[0053] mσ1 is taken as the target value for the brightness determination of condensation in the cold mirror dew point meter. Based on the experiment, m=6 is determined, and an allowable fluctuation range of the brightness L of the detection image is set according to the maximum and minimum values ​​of the measured data.

[0054] Step S4: Cool the air chamber for the first time, start the dynamic PID controller, and adjust the parameters of the dynamic PID controller according to the deviation between the brightness L of the real-time acquired detection image and the target value 6σ1 of the brightness judgment value; when the deviation is large, increase the proportional coefficient, decrease the integral time constant, and increase the derivative time constant to improve the response speed and stability of temperature control; when the deviation is small, do the opposite.

[0055] Step S5: When the temperature inside the air chamber decreases until the brightness of the detected image reaches the target value of the brightness judgment value, i.e., L = 6σ1, the area prone to condensation is identified, and the cooling of the air chamber is stopped; when the temperature inside the air chamber decreases until L = 6σ1, the current detected image is processed by median filtering, a 2x2 filter kernel is selected to remove interference from two or fewer pixels, and the scattered dew points are marked as areas prone to condensation. The cold mirror dew point meter immediately stops cooling.

[0056] Step S6: Mark all pixels in the condensation-prone areas of the detection image as 1 and the rest as 0, and call it the template image. Perform a bitwise AND operation between the detection image and the template image to obtain a detection image that retains only the condensation-prone areas, and call it the ROI region detection image.

[0057] Step S7: Measure and calculate the brightness change caused by detection error in the ROI region detection image when there is no condensation. Calculate its variance σ2 multiple times and redetermine the target value of brightness judgment value 6σ2 as the brightness threshold value for judging condensation.

[0058] Step S8: Cool the air chamber a second time, restart the dynamic PID controller, adjust the parameters of the dynamic PID controller according to the deviation between the real-time detected image brightness L and the brightness judgment value target value 6σ2, calculate the brightness F of the detected image of the ROI area, and stop cooling immediately when the system cools down to the brightness judgment value target value, i.e. F=6σ2, and record the temperature at this time as the dew point temperature T1.

[0059] Step S9: The temperature of the air chamber gradually increases, and the dew point gradually disappears. When the temperature inside the air chamber rises to F = σ2, record the temperature at this time as the dew point temperature T2.

[0060] Step S10: Calculate the dew point temperature by combining the dew point temperatures T1 and T2 with a set weight. Specifically, fuse the read dew point temperatures T1 and T2 separately to calculate the measured dew point temperature value. The calculation formula is as follows:

[0061] T = kT1 + (1-k)T2

[0062] Where k is the weighting coefficient, and k = 0.7 was determined based on experiments;

[0063] The test is complete.

[0064] The aforementioned meteorological detection method provides an image recognition method that can improve detection accuracy and reliability. To facilitate the smooth implementation of this method, a cold mirror dew point meter is needed. Its main purpose is to determine whether condensation has formed on the mirror surface based on changes in pixel brightness when capturing images of the mirror surface using a camera. For example... Figure 1-3 The figure shows a preferred embodiment of this application, including a housing for the main control room. A cooling stack is positioned at the center of the top of the housing. As shown in the figure, the cooling stack comprises several sequentially stacked cooling elements, with the cooling element at the top of the stack having a smaller area than the others in the vertical direction. A gas chamber is located at the top of the cooling stack, and a cold mirror is embedded in the gas chamber relative to the bottom of the stack. The cooling stack is used to control the mirror's surface temperature. A platinum resistance thermometer 10 (PT1000, four-wire) is embedded in the center of the cold mirror for reading the mirror's surface temperature. A camera is mounted on top of the gas chamber, with its lens facing the center of the cold mirror. The camera's focus can be freely adjusted to ensure clear imaging, and it is used to capture images of the mirror surface. An air inlet 11 (for introducing the gas to be tested into the chamber) and an air outlet 3 are respectively set on both sides of the gas chamber. It is important to note that the gas chamber, except for the air inlet and outlet, is a sealed structure with an internal black dye coating for high light-blocking properties. A light-emitting diode (LED) is also installed on the side wall where the air outlet is located. The light emitted by the LED is directed towards the cold mirror surface at a set incident angle. Setting the incident angle ensures that the incident light can reach the cold mirror surface, laying the groundwork for subsequent work. After multiple experiments, it was found that a 50° incident angle resulted in the optimal light coverage area. A heat dissipation component 8, combining air and water cooling, is installed at the bottom of the enclosure to dissipate heat from the refrigeration unit. Since the technology for heat dissipation components is relatively mature, no specific selection is made here. A touchscreen 5 is installed on one side of the enclosure; the main control system is connected to the refrigeration unit, camera, touchscreen, and heat dissipation component.

[0065] As can be seen from the description of this application, when this application is in operation, the main control room controls the camera to collect the mirror image of the cold mirror and calculate the image brightness. First, the area prone to condensation is located, and then the brightness of the area prone to condensation is detected. Furthermore, by fusing the temperature at which condensation occurs and the temperature at which the dew point disappears, the condensation status of the mirror can be accurately determined, which greatly improves the accuracy and reliability of dew point detection.

[0066] Those skilled in the art will understand that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0067] The meaning of "and / or" as used in this application includes both situations where each exists alone or both exist simultaneously.

[0068] The term "connection" as used in this application can mean a direct connection between components or an indirect connection between components through other components.

[0069] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A meteorological detection method based on image recognition, characterized in that: Specifically, the following steps are included: Step S1: Start the cold mirror dew point meter. High-pressure gas is sprayed onto the cold mirror surface for cleaning. After cleaning is completed, the refrigeration unit applies a reverse voltage to the bottom of the cold mirror to heat it until the cold mirror surface is dry. Step S2: Simultaneously activate the LED and camera, input the gas to be tested into the gas chamber, the camera acquires real-time mirror images of the cold mirror, and sends the acquired mirror images to the main control room; select the detection image from several mirror images; Step S3: Take multiple measurements and, based on the selected detection images, calculate the variance σ1 of the brightness change caused by detection error when the cold mirror surface is free of condensation; Step S4: Cool the air chamber for the first time, start the dynamic PID controller, and adjust the parameters of the dynamic PID controller according to the deviation between the real-time acquired detection image brightness L and the brightness judgment target value mσ1. Step S5: When the temperature inside the air chamber decreases until the brightness of the detected image reaches the target value of the brightness judgment value, i.e., L=mσ1, the area prone to condensation is identified, and the cooling of the air chamber is stopped; In step S5, when the temperature inside the air chamber decreases until L = 6σ1, the current detection image is processed by median filtering. A 2x2 filter kernel is selected to remove interference from two or fewer pixels, and scattered dew points are identified as areas prone to condensation. The cold mirror dew point meter immediately stops cooling. Step S6: Mark all pixels in the condensation-prone areas of the detection image as 1 and the rest as 0, and call it the template image. Perform a bitwise AND operation between the detection image and the template image to obtain a detection image that retains only the condensation-prone areas, and call it the ROI region detection image. Step S7: Measure and calculate the brightness change caused by detection error in the ROI region detection image when there is no condensation. Calculate its variance σ2 multiple times and redetermine the target value mσ2 of the brightness judgment value as the brightness threshold value for judging condensation. Step S8: Cool the air chamber a second time, restart the dynamic PID controller, adjust the parameters of the dynamic PID controller according to the deviation between the real-time detected image brightness L and the brightness judgment target value mσ2, calculate the brightness F of the detected image of the ROI area, and stop cooling immediately when the system cools down to the brightness judgment target value, i.e. F=mσ2, and record the temperature at this time as the dew point temperature T1. Step S9: The temperature of the air chamber gradually increases, and the dew point gradually disappears. When the temperature inside the air chamber rises to F = σ2, record the temperature at this time as the dew point temperature T2. Step S10: Calculate the dew point temperature by using the dew point temperature T1 and dew point temperature T2 with the set weights, and the detection ends.

2. The meteorological detection method based on image recognition according to claim 1, characterized in that: In step S2, the 10 frames of images taken when the cold mirror surface is free of condensation are summed and averaged to obtain the background image. The subsequent 10 frames of images are summed and averaged, and the background image is subtracted to obtain the detection image.

3. The meteorological detection method based on image recognition according to claim 1, characterized in that: In step S3, multiple measurements are taken, and the variance σ1 of the brightness change caused by detection error when there is no condensation is calculated. The calculation formula is as follows: ； Where L is the brightness of the detected image acquired in real time, L=R*0.299+G*0.587+B*0.114, R, G, B are the values ​​of the red, green and blue color channels in the RGB color space, respectively; The value is the average brightness variation; n is the number of measurements. mσ1 is taken as the target value for the brightness determination of condensation in the cold mirror dew point meter. Based on the experiment, m=6 is determined, and an allowable fluctuation range of the brightness L of the detection image is set according to the maximum and minimum values ​​of the measured data.

4. The meteorological detection method based on image recognition according to claim 1, characterized in that: In step S4, when the deviation is large, the proportional coefficient is increased, the integral time constant is decreased, and the derivative time constant is increased to improve the response speed and stability of temperature control; when the deviation is small, the opposite is true.

5. The meteorological detection method based on image recognition according to claim 1, characterized in that: In step S10, the dew point temperatures T1 and T2, read separately, are fused to calculate the measured dew point temperature value. The calculation formula is as follows: ； Where k is the weighting coefficient, and k = 0.7 was determined based on experiments.

Images