Method for detecting waste heat transfer efficiency based on infrared image

Abstract

The present invention relates to the field of waste heat transfer, and more specifically, to a method for detecting the waste heat transfer efficiency based on infrared images. The method includes: obtaining the ambient temperature, ambient humidity, and infrared images of each band at any sampling moment of the pipeline, and calculating the noise level of each band; for any band, obtaining the actual variance and actual convolution kernel according to the noise level, and filtering the infrared image with the actual variance and actual convolution kernel to obtain a filtered infrared image; traversing to obtain the filtered infrared images of each band, using the noise level as a weight, and performing weighted summation on the filtered infrared images of each band to obtain a fused infrared image, thereby completing the detection of the transfer efficiency. Through the technical solution of the present invention, the accuracy of the waste heat transfer efficiency detection result can be improved.

Description

Technical Field

[0001] The present invention relates to the field of waste heat transfer. More specifically, the present invention relates to a method for detecting the waste heat transfer efficiency based on infrared images. Background Art

[0002] The problem of energy consumption has increasingly become a global focus. During the process of energy utilization, the recovery and utilization of waste heat can not only significantly improve the energy use efficiency, but also play a crucial role in reducing greenhouse gas emissions, lowering production costs, protecting the environment, etc. The waste heat utilization technology is particularly widely used in fields such as metallurgy, chemical industry, machinery manufacturing, power generation, etc., and is one of the key technologies to improve industrial energy-saving efficiency.

[0003] Traditional waste heat detection technologies mostly rely on devices such as sensors and thermometers. Common methods include temperature measurement, flow monitoring, etc. However, in the face of complex industrial environments and high-speed heat transfer processes, these traditional methods often face problems such as large measurement errors or incomplete data. Especially in some high-temperature or inaccessible areas, the deployment and monitoring of traditional sensors have great limitations. Due to the advantages of non-contact measurement and real-time monitoring of infrared imaging technology, it has gradually become an emerging means for detecting the waste heat transfer efficiency. Through an infrared thermal imager, the thermal radiation information of the target object can be obtained, and thus the temperature distribution can be inferred.

[0004] However, due to the presence of noise interference in different infrared images, simply analyzing the images in a single dimension often cannot comprehensively reflect the true situation of waste heat transfer, which may lead to distortion of temperature data; noise will not only cause deviations in temperature measurement, affecting the accuracy of measurement results, but may also mask the subtle changes in the heat transfer process, resulting in inaccurate detection results of waste heat transfer efficiency. Summary of the Invention

[0005] To solve the problem of inaccurate detection results of waste heat transfer efficiency, the present invention proposes a method for detecting the waste heat transfer efficiency based on infrared images.

[0006] The present invention discloses a method for detecting the waste heat transfer efficiency based on infrared images, including: obtaining the ambient temperature, ambient humidity, and infrared images of each band at any sampling moment of the pipeline, calculating the noise level of each band, where one sampling moment corresponds to several bands; for any band, obtaining the actual variance and actual convolution kernel according to the noise level, and filtering the infrared image with the actual variance and actual convolution kernel to obtain a filtered infrared image; traversing to obtain the filtered infrared images of each band, using the noise level as the weight, and performing weighted summation on the filtered infrared images of each band to obtain a fused infrared image, and completing the detection of the transfer efficiency.

[0007] By combining the ambient temperature, humidity, and the noise levels of infrared images in each band, the filtering method for each band is intelligently adjusted, thereby optimizing the image processing effect. Appropriate filtering parameters are dynamically selected according to the noise characteristics of the band to precisely process the infrared image, reducing noise interference and effectively retaining useful information. By weighted fusion of the filtered images in different bands, not only can the image quality be improved and details enhanced, but also appropriate weights can be assigned to each band according to the noise level to ensure the clarity and information integrity of the fused image.

[0008] Preferably, the noise level includes: obtaining the central temperature and central humidity of any band in history; the noise level satisfies the relational expression:

[0009] , represents the noise level, and respectively represent the ambient temperature and ambient humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the natural constant.

[0010] It can dynamically adjust the image processing strategy according to the relative change between the current environmental conditions and the historical environmental state, ensuring the optimization of the noise suppression effect under different environmental conditions. This method improves the noise adaptability in the infrared image processing process.

[0011] Preferably, the noise level includes: obtaining the credible temperature range and credible humidity range of any band in history, and obtaining the central temperature and central humidity; the noise level satisfies the relational expression:

[0012] , represents the noise level, and respectively represent the ambient temperature and ambient humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the credible temperature range, represents the credible humidity range, represents the natural constant.

[0013] By introducing the credible temperature range and credible humidity range, it is ensured that the sensitivity to noise can be flexibly adjusted under different environmental conditions. The adaptability to environmental changes is effectively improved, enabling the noise impact to be accurately quantified and controlled, so that unnecessary noise interference can be reduced and more effective information can be retained during the filtering and fusion processes.

[0014] Preferably, the central temperature and central humidity include: obtaining all historical infrared images in any wavelength band in history, adding labels to each historical infrared image, where the labels are qualified and unqualified; taking the central point of the ambient temperature of all historical infrared images with qualified labels as the central temperature, and taking the central point of the ambient humidity of all historical infrared images with qualified labels as the central humidity.

[0015] Preferably, the central temperature and central humidity include: obtaining all historical infrared images in any wavelength band in history, adding labels to each historical infrared image, where the labels are qualified and unqualified; taking the mode of the ambient temperature of all historical infrared images with qualified labels as the central temperature, and taking the mode of the ambient humidity of all historical infrared images with qualified labels as the central humidity.

[0016] Preferably, the credible temperature range and credible humidity range include: taking the value range of the ambient temperature of all historical infrared images with qualified labels as the credible temperature range, and taking the value range of the ambient humidity of all historical infrared images with qualified labels as the credible humidity range.

[0017] Preferably, the obtaining of the actual variance and actual convolution kernel includes: presetting a filtering variance and a filtering convolution kernel; taking the product of the noise level of any wavelength band and the filtering variance as the actual variance, and taking the product of the noise level of any wavelength band and the filtering convolution kernel as the actual convolution kernel.

[0018] The introduction of the noise level enables the filtering parameters to be flexibly adjusted according to the current environment, so as to process wavelength bands with different noise intensities more reasonably. Specifically, wavelength bands with a higher noise level will use smaller actual variances and convolution kernels to reduce the impact of noise on the image, while wavelength bands with less noise can use larger parameters to better retain the details of the image.

[0019] Preferably, the obtaining of the actual variance and actual convolution kernel includes: presetting a filtering variance, and taking the product of the noise level of any wavelength band and the filtering variance as the actual variance; in response to the actual variance expanding by a preset multiple, the size of the actual convolution kernel is expanded once.

[0020] Preferably, the obtained fused infrared image satisfies the relation:

[0021] , represents the fused infrared image, represents the wavelength band of the noise level, represents the wavelength band of the filtered infrared image, represents the natural constant.

[0022] By introducing an exponential decay function of the noise level and assigning different weights to each band, the fused image can more accurately reflect the effective information of each band and suppress unnecessary noise interference. This weighted fusion method effectively improves the quality of the final image, ensures the retention of more important information, and reduces the distortion introduced by noise.

[0023] Preferably, the detection of the transmission efficiency includes: performing threshold segmentation on the fused infrared image to obtain the waste heat transmission area, calculating the average pixel value of the waste heat transmission area, obtaining the temperature of the waste heat transmission area according to the mapping relationship between the pixel value and the actual temperature in the infrared image, calculating the actual heat transfer amount according to the temperature of the waste heat transmission area, and taking the ratio of the actual heat transfer amount to the theoretical heat transfer amount as the transmission efficiency to complete the detection of the transmission efficiency.

[0024] Advantages of the present invention:

[0025] By combining the ambient temperature and humidity changes and image processing technology, the present invention can more accurately evaluate the waste heat transmission status in the pipeline. Using infrared image processing technology, the waste heat transmission area can be effectively extracted and analyzed. By processing the noise of the infrared image, background interference is reduced and irrelevant factors are eliminated, thereby improving the clarity and accuracy of the image. The weighted fusion technology can combine image information from different angles to obtain more comprehensive and accurate temperature distribution data. Threshold segmentation helps to distinguish different temperature regions and accurately capture the waste heat transmission area.

[0026] The present invention can more accurately identify the heat transfer path, thereby providing a more accurate calculation of the heat transfer amount, optimizing the energy utilization efficiency and effectively reducing energy loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flowchart of the method for detecting the waste heat transmission efficiency based on the infrared image according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention.

[0029] The following will describe the specific embodiments of the present invention in detail with reference to the accompanying drawings.

[0030] Refer to Figure 1 , the method for detecting the waste heat transmission efficiency based on the infrared image includes steps S1 - S3, which will be specifically described below.

[0031] S1. Obtain the ambient temperature, ambient humidity, and infrared images of each band at any sampling moment of the pipeline, and calculate the noise level of each band, where one sampling moment corresponds to several bands.

[0032] In one embodiment, install a temperature sensor and a humidity sensor near the pipeline responsible for waste heat transfer to detect the ambient temperature and humidity in real time, and ensure that the acquisition of ambient temperature and humidity is synchronized, that is, start and end simultaneously.

[0033] Install an infrared device to obtain several infrared images at any sampling moment. It should be noted that when collecting infrared images, infrared images of different bands should be collected at the same sampling moment, that is, one sampling moment corresponds to multiple bands, and each band corresponds to one infrared image. This is because infrared light of different bands exhibits different acquisition effects under different environmental conditions. For example, mid-wave infrared light can obtain higher-quality images in high-temperature and high-humidity environments, while infrared light of other bands is more susceptible to more noise interference in such environments, resulting in poor image quality. Therefore, collecting ambient temperature and humidity data can help judge the quality of infrared images of different bands and provide more accurate environmental background information for subsequent analysis.

[0034] Obtain the central temperature and central humidity of any band in history, including obtaining all historical infrared images of any band in history, and adding labels to each historical infrared image, with the labels being qualified and unqualified. The way to obtain the labels is as follows: Those skilled in the art evaluate the shooting quality of each historical infrared image in combination with the actual situation, and use the evaluation result as the label, and the evaluation result includes qualified and unqualified.

[0035] Take the central point of the ambient temperature of all historical infrared images with qualified labels as the central temperature, and take the central point of the ambient humidity of all historical infrared images with qualified labels as the central humidity.

[0036] The noise level satisfies the relationship: , represents the noise level, and respectively represent the ambient temperature and ambient humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the natural constant.

[0037] It should be noted that at the same sampling moment, the environmental temperature and humidity remain constant, that is, the environmental temperature and humidity corresponding to all bands at this moment are the same. However, the infrared images of different bands have different adaptation ranges to the environmental temperature and humidity, which are called the credible temperature range and credible humidity range of this band. Each band has an optimal environmental temperature and humidity value in the historical data, which is called the central temperature and central humidity of this band. The central temperature and central humidity of each band are different, and there are differences between these values among different bands.

[0038] To evaluate the quality of the infrared image of a band, the difference between the current environmental temperature and the central temperature of this band, as well as the difference between the current environmental humidity and the central humidity of this band, can be compared. If the temperature and humidity of the current environment deviate less from the central values of this band, it indicates that the current environmental conditions are suitable for the infrared imaging of this band, and the quality and credibility of the infrared image are relatively high. On the contrary, if the difference is large, it may lead to poor image quality and reduced credibility.

[0039] In one embodiment, the noise level includes:

[0040] Obtain the credible temperature range and credible humidity range of any band in history. The credible temperature range and credible humidity range include: taking the value range of the environmental temperature of all historical infrared images labeled as qualified as the credible temperature range, and taking the value range of the environmental humidity of all historical infrared images labeled as qualified as the credible humidity range.

[0041] Obtain the central temperature and central humidity.

[0042] The noise level satisfies the relational expression: , represents the noise level, and respectively represent the environmental temperature and environmental humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the credible temperature range, represents the credible humidity range, represents the natural constant.

[0043] The farther the environmental temperature and environmental humidity at the current sampling moment are from the optimal adaptation environment of a certain band, the more likely it is to cause unstable quality of the infrared image of this band and increased noise. This is because the infrared image of each band has an optimal response interval for specific environmental conditions. If it deviates from this interval, the signal-to-noise ratio of the image will decrease, resulting in more obvious noise, thereby affecting the reliability of the infrared image.

[0044] In one embodiment, the central temperature and central humidity include: obtaining all historical infrared images of any wavelength in history, adding labels to each historical infrared image, where the labels are qualified and unqualified; taking the mode of the ambient temperatures of all historical infrared images with qualified labels as the central temperature, and taking the mode of the ambient humidities of all historical infrared images with qualified labels as the central humidity.

[0045] S2. For any wavelength, obtain the actual variance and the actual convolution kernel according to the noise level, and use the actual variance and the actual convolution kernel to filter the infrared image to obtain a filtered infrared image.

[0046] In one embodiment, Gaussian filtering is a commonly used image processing technique mainly used to remove noise and smooth images. In Gaussian filtering, the convolution kernel and the variance are two key parameters, where the variance directly affects the calculation of the standard deviation, and the standard deviation determines the smoothness of the Gaussian kernel. Specifically, the larger the standard deviation, the stronger the smoothing effect of the Gaussian kernel, and the details of the image will be further blurred or lost. In this way, Gaussian filtering can effectively remove high-frequency noise, but it may also cause the loss of details in the image, especially when the image contains important texture or edge information. Therefore, the selection of the variance needs to be adjusted according to the actual situation to balance the relationship between noise removal and detail retention. If the variance is set too large, the image will become too blurred and lose too many details; if the variance is too small, it may not be able to effectively remove noise and affect the clarity of the image.

[0047] Preset the filtering variance and the filtering convolution kernel; take the product of the noise level of any wavelength and the filtering variance as the actual variance, and take the product of the noise level of any wavelength and the filtering convolution kernel as the actual convolution kernel.

[0048] Filter the infrared image of any wavelength according to the true variance and the true convolution kernel to obtain a filtered infrared image.

[0049] In one embodiment, obtaining the actual variance and the actual convolution kernel includes:

[0050] Preset the filtering variance, and take the product of the noise level of any wavelength and the filtering variance as the actual variance; in response to the actual variance expanding by a preset multiple, the size of the actual convolution kernel expands once.

[0051] Exemplarily, the variance values can be 1, 2, 3, 4, etc. from small to large until positive infinity, and the size of the convolution kernel is generally 3 by 3, 5 by 5, 7 by 7, etc. Therefore, there is a mapping relationship between the variance and the convolution kernel, that is, when the variance is 1, 2, 3, the size of the convolution kernel is 3 by 3, and when the variance is 4 to 12, the size of the convolution kernel is 5 by 5, and so on, to obtain the convolution kernel sizes corresponding to different variances. Thus, after obtaining the true variance, the true convolution kernel can be correspondingly obtained.

[0052] S3. Traverse to obtain the filtered infrared images of each band, use the noise level as the weight, and perform weighted summation on the filtered infrared images of each band to obtain a fused infrared image, thus completing the transmission efficiency detection.

[0053] In one embodiment, obtaining the fused infrared image satisfies the relation:

[0054] , represents the fused infrared image, represents the band of the noise level, represents the band of the filtered infrared image, represents the natural constant.

[0055] Bands with a higher noise level will be given a smaller weight, while bands with a lower noise level will be given a larger weight, so as to preferentially retain the band data with clearer signals and less noise when fusing images.

[0056] Perform threshold segmentation on the fused infrared image to obtain the waste heat transfer area, calculate the average pixel value of the waste heat transfer area, obtain the temperature of the waste heat transfer area according to the mapping relationship between the pixel value and the actual temperature in the infrared image, calculate the actual heat transfer amount according to the temperature of the waste heat transfer area, and use the ratio of the actual heat transfer amount to the theoretical heat transfer amount as the transmission efficiency, thus completing the transmission efficiency detection.

[0057] It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for detecting the waste heat transfer efficiency based on infrared images, characterized in that, Including: Obtain the ambient temperature, ambient humidity, and infrared images of each band at any sampling moment of the pipeline, ensuring that the acquisition of ambient temperature and humidity is synchronized so that the ambient temperature and humidity corresponding to all bands at this moment are the same. Calculate the noise level of each band. One sampling moment corresponds to several bands. For any band, obtain the actual variance and actual convolution kernel according to the noise level, and use the actual variance and actual convolution kernel to filter the infrared image to obtain a filtered infrared image. Traverse to obtain the filtered infrared images of each band, use the noise level as the weight, and perform weighted summation on the filtered infrared images of each band to obtain a fused infrared image, and complete the transmission efficiency detection. The noise level includes: Obtain the central temperature and central humidity of any band in history. The noise level satisfies the relation: , represents the noise level, and respectively represent the ambient temperature and ambient humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the natural constant; The obtaining of the actual variance and actual convolution kernel includes: Preset the filtering variance and filtering convolution kernel. Take the product of the noise level of any band and the filtering variance as the actual variance, and take the product of the noise level of any band and the filtering convolution kernel as the actual convolution kernel.

2. The method for detecting the waste heat transfer efficiency based on infrared images according to claim 1, wherein The noise level also includes: Obtain the credible temperature range and credible humidity range of any band in history, and obtain the central temperature and central humidity. The noise level satisfies the relation: , represents the noise level, and respectively represent the ambient temperature and ambient humidity at any sampling moment, represents the central temperature, represents the central humidity, represents the credible temperature range, represents the credible humidity range, represents the natural constant.

3. The method for detecting the waste heat transfer efficiency based on infrared images according to claim 2, wherein The central temperature and central humidity include: Obtain all historical infrared images of any band in history, and add labels to each historical infrared image, with the labels being qualified and unqualified. Take the central point of the ambient temperature of all historical infrared images with the label of qualified as the central temperature, and take the central point of the ambient humidity of all historical infrared images with the label of qualified as the central humidity.

4. The method for detecting the waste heat transfer efficiency based on infrared images according to claim 2, wherein The central temperature and central humidity include: Obtain all historical infrared images of any band in history, and add labels to each historical infrared image, with the labels being qualified and unqualified. Take the mode of the ambient temperature of all historical infrared images with the label of qualified as the central temperature, and take the mode of the ambient humidity of all historical infrared images with the label of qualified as the central humidity.

5. The method for detecting the waste heat transfer efficiency based on infrared images according to claim 2, wherein The credible temperature range and credible humidity range include: Take the value range of the ambient temperature of all historical infrared images with the label of qualified as the credible temperature range, and take the value range of the ambient humidity of all historical infrared images with the label of qualified as the credible humidity range.

6. The method for detecting the waste heat transfer efficiency based on an infrared image according to claim 1, wherein The obtaining of the actual variance and actual convolution kernel also includes: Preset the filtering variance, and take the product of the noise level of any band and the filtering variance as the actual variance. In response to the actual variance expanding by a preset multiple, the size of the actual convolution kernel is expanded once.

7. The method for detecting the waste heat transfer efficiency based on infrared images according to claim 1, characterized in that The obtained fused infrared image satisfies the relation: , represents the fused infrared image, represents the waveband of the noise level, represents the waveband of the filtered infrared image, represents the natural constant.

8. The method for detecting the waste heat transfer efficiency based on an infrared image according to claim 1, wherein The completion of the transmission efficiency detection includes: Perform threshold segmentation on the fused infrared image to obtain the waste heat transmission area, calculate the average pixel value of the waste heat transmission area, obtain the temperature of the waste heat transmission area according to the mapping relationship between the pixel value and the actual temperature in the infrared image, calculate the actual heat transfer according to the temperature of the waste heat transmission area, and take the ratio of the actual heat transfer to the theoretical heat transfer as the transmission efficiency to complete the transmission efficiency detection.

Images