Offshore platform equipment thermal infrared imaging detection early warning method based on mobile terminal

Abstract

The invention relates to an offshore platform equipment thermal infrared imaging detection early warning method based on a mobile terminal, and the method comprises the steps: enabling the mobile terminal to be connected with a thermal infrared camera, and inputting the information of target detection equipment; the thermal infrared camera extracts the equipment temperature in the thermal infrared image by adopting a thermal imaging algorithm and uploads the equipment temperature to the server; the server extracts the highest temperature, the lowest temperature and the average temperature in the image by using a character recognition algorithm, and stores the temperatures in a database; a vibration sensor and a temperature and humidity sensor installed on the equipment collect vibration displacement data, vibration frequency data, temperature and humidity data in real time, upload the data to a server and store the data in a database; and when more than one piece of data acquired and extracted in real time continuously exceeds a set corresponding threshold within a specific time duration, the server judges that the equipment is in an abnormal state and feeds back early warning information to the mobile terminal. The method is suitable for fault detection of complex environment equipment of an offshore platform, and faults and accidents caused by high temperature, vibration or gas leakage and the like are prevented.

Description

Technical Field

[0001] The present invention relates to a thermal infrared imaging detection and early warning method for offshore platform equipment based on a mobile terminal, which is suitable for inspection and fault early warning of target detection equipment in a complex environment of an offshore platform. Background Art

[0002] Offshore platforms are equipped with many important equipment for marine oil and gas development. The safe and stable operation of the equipment is an important factor in ensuring production efficiency and reducing unexpected failures. The operating environment of offshore platform equipment is complex and is often affected by marine climate, temperature fluctuations, vibrations and other harsh conditions, which makes the equipment prone to overheating, abnormal vibrations or gas leaks. If these abnormal conditions are not discovered in time, they may cause serious equipment damage or safety accidents. Existing offshore platform equipment monitoring methods mostly rely on regular manual inspections or data analysis of a single type of sensor. It is difficult to detect potential abnormal conditions of equipment in real time, and the identification of hidden dangers such as high temperature and vibration is not timely enough. Summary of the invention

[0003] The purpose of the present invention is to provide a thermal infrared imaging detection and early warning method for offshore platform equipment based on a mobile terminal. By combining the mobile terminal with a thermal infrared camera for inspection, the temperature of the equipment can be monitored in real time. Combined with the data from the vibration sensor and the temperature and humidity sensor, real-time monitoring of abnormal conditions of the equipment can be achieved, thereby improving the safety and stability of the offshore platform equipment.

[0004] The implementation steps of the method include: first, the mobile terminal is connected to the thermal infrared camera, and the area and device name of the offshore platform target detection device are input through the mobile terminal. The thermal infrared camera uses a thermal imaging algorithm to obtain the thermal infrared image of the device and uploads the image to the server. The server automatically extracts the highest temperature, lowest temperature and average temperature data in the image through a text recognition algorithm, and stores the data in a database. The vibration sensor and temperature and humidity sensor installed on the device collect vibration displacement data, vibration frequency data, temperature and humidity data in real time and upload them to the server. If at least one of the vibration displacement data, vibration frequency data, temperature and humidity data of the target device, highest temperature, lowest temperature and average temperature data exceeds the corresponding data threshold set in advance for more than a specified specific time length, the server will determine that the device is in an abnormal working state and feedback the warning information to the mobile terminal. All data can be viewed, edited and managed through the constructed web page, and users can adjust the threshold or update the device information as needed.

[0005] The thermal infrared camera is a thermal imaging device with a Type-C interface, which can be connected to a mobile device and can be equipped with a thermal imaging algorithm model to generate corresponding thermal infrared images. The thermal imaging processing algorithm invented includes:

[0006] S201, noise filtering and temperature region segmentation are performed on the thermal infrared image. In order to improve the image analysis accuracy of the offshore platform and eliminate the error caused by the environment or equipment noise, an improved Gaussian filter is used; the improved Gaussian filter introduces a weight w(i, j) parameter to adjust the influence of different pixels in the filtering process. Each pixel value I′(x, y) after filtering can be calculated by the following formula;

[0007]

[0008] Among them, I(x,y) is the pixel value of the original image, G(i,j) is the two-dimensional Gaussian kernel function, k is the size of the filter window, W is the normalization factor to ensure that the pixel value after filtering remains within a reasonable range, and w(i,j) is the weight function, which is dynamically adjusted according to the noise characteristics of the local pixel to adapt to the complex situation of the offshore platform;

[0009] S202, through image pixel-level temperature calculation, identify the hot spot area on the surface of the equipment, generate the corresponding thermal distribution map, and convert the infrared radiation intensity into temperature based on the principle of thermal imaging and the particularity of the offshore platform environment. The specific formula is as follows:

[0010]

[0011] Where L is the infrared radiation intensity and σ is the Stefan-Boltzmann constant 5.67×10 -8 W / m 2 K 4 , T is the absolute temperature in Kelvin, ∈ is the emissivity of the device surface, and the emissivity of the device surface is 0.95; C is the correction factor, which is 0.9 at the upper temperature limit and 1.1 at the lower temperature limit to ensure that the temperature of thermal infrared recognition is within this range; by calculating each pixel point in the image, the radiation intensity of each pixel is converted into the corresponding temperature value.

[0012] The text recognition algorithm recognizes the temperature detected by thermal imaging. The invented text recognition algorithm includes:

[0013] S301, before performing text recognition, firstly perform contrast enhancement and image sharpening processing on the text area in the thermal infrared image, and improve the text clarity in the thermal infrared image and increase the recognizability of the text by using the contrast adaptive histogram equalization (CLAHE) algorithm. The CLAHE algorithm formula is as follows:

[0014]

[0015] Among them, CDF(I(x,y)) is the cumulative distribution function of the pixel value I(x,y), M×N is the total number of pixels in the image, L is the gray level, and CDF min is the minimum non-zero value in the cumulative distribution function. The text part in the image is enhanced by this algorithm to make it clearer;

[0016] S302, identifying the area containing text in the thermal infrared image through the DBNet text detection model, effectively segmenting the text part in the thermal infrared image, and using the DBNet algorithm for text area detection to generate the text area through the binarization probability map;

[0017] S303: For the detected text area, a text recognition model is used for recognition. The model adopts a convolutional recurrent neural network (CRNN) structure, extracts image features through multi-layer convolution, and uses a bidirectional LSTM with an attention mechanism to perform character-level sequence decoding, so that the network can better capture the dependency between contexts and finally output the text content. The formula of the bidirectional LSTM with an attention mechanism is as follows:

[0018] h t =Attention(x t ,h t-1 )·σ(W h ·x t +U h ·h t-1 +b h )

[0019] Among them, x t is the input feature vector, h t is the hidden state at the current moment, W h and U h is the weight matrix, b h is the bias term, Attention(x t ,h t-1 ) is an attention mechanism that calculates the importance of each character based on the current input and historical hidden states. CRNN realizes sequence-to-sequence conversion through this model and finally outputs the highest temperature, lowest temperature and average temperature data in the recognized thermal infrared image.

[0020] The highest, lowest and average temperatures extracted by the text recognition algorithm and the vibration displacement data, vibration frequency data, temperature data and humidity data collected by the vibration sensors and temperature and humidity sensors installed on the equipment are uploaded to the server and stored in the database. When at least one data exceeds the corresponding data threshold set in advance for more than a specified specific time length, the server will determine that the equipment is in an abnormal working state, give a warning message, and feed back the results to the mobile terminal.

[0021] The server can extract relevant data from the database and display it on the web page. Users can access it through the browser and view the temperature data and warning information of the device in real time. The warning information of the device is transmitted to the mobile device in real time through Socket communication and displayed on the mobile page.

[0022] The beneficial effects of the present invention are as follows:

[0023] 1. Real-time monitoring: The present invention realizes real-time monitoring of offshore platform equipment through the combination of a mobile terminal and a thermal infrared camera, thereby improving the convenience of monitoring.

[0024] 2. Thermal imaging analysis and text recognition and extraction: The thermal imaging algorithm is used to analyze the infrared radiation intensity of each pixel to generate an accurate temperature distribution map. The server uses a text recognition algorithm to accurately identify and extract the highest temperature, lowest temperature and average temperature data in the thermal infrared image, and stores the data to ensure data reliability.

[0025] 3. Multi-sensor data fusion: The present invention combines vibration sensors and temperature and humidity sensors to obtain the vibration displacement, vibration frequency, temperature and humidity data of the target equipment. By comprehensively analyzing these data with the highest temperature, lowest temperature and average temperature obtained by the thermal infrared camera, the status of the equipment can be fully monitored.

[0026] 4. Abnormal detection and early warning: When at least one of the vibration displacement data, vibration frequency data, temperature and humidity data, maximum temperature, minimum temperature and average temperature data exceeds the preset threshold within a specified time period, the server will determine that the device is in an abnormal working state and give an early warning message. This improves the accuracy of fault detection and reduces the possibility of false alarms and missed alarms.

[0027] 5. Web page data management: The web page of the present invention provides comprehensive data management functions. Users can view and edit the collected data, historical records and warning information in real time through the web page, thereby achieving more convenient and comprehensive equipment management. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of thermal infrared imaging detection and early warning of offshore platform equipment based on a mobile terminal according to an embodiment of the present invention.

[0029] Figure 2 Schematic diagram of the thermal imaging algorithm structure constructed according to the present invention.

[0030] Figure 3 It is a schematic diagram of the structure of a text recognition algorithm constructed according to the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be described clearly and completely in conjunction with the drawings in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0032] With the development of monitoring technology, thermal infrared imaging technology can monitor the temperature changes of equipment in real time and effectively detect equipment overheating problems. Vibration sensors and environmental monitoring sensors can provide vibration and temperature and humidity data of equipment operation, providing a comprehensive reference for the equipment operation status. The present invention improves the accuracy and real-time performance of offshore platform equipment fault detection by combining thermal imaging technology, text recognition technology, and sensor detection technology.

[0033] The present invention proposes a thermal infrared imaging detection and early warning method for offshore platform equipment based on a mobile terminal. Figure 1 This is a flow chart of the thermal infrared imaging detection and early warning method for offshore platform equipment based on a mobile terminal. The process includes the following six steps:

[0034] S100, the mobile terminal is connected to a thermal infrared camera, and the mobile terminal receives the location and name of the offshore platform target detection device input by a user;

[0035] S200, photographing an image of a target detection device by means of the thermal infrared camera, extracting a temperature distribution area of ​​the target detection device by means of a thermal imaging algorithm, extracting temperature data from the thermal infrared image, and uploading the data to a server;

[0036] S300, the server uses a text recognition algorithm to recognize the thermal infrared image, extracts the maximum temperature, minimum temperature and average temperature data in the thermal infrared image, and stores the maximum temperature, minimum temperature and average temperature data in a database;

[0037] S400, obtaining vibration displacement data and vibration frequency data of the target device through the vibration sensor installed on the target detection device, obtaining temperature and humidity data of the target detection device through the temperature and humidity sensor installed on the target detection device, and uploading the collected thermal infrared image, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection device to the server, and storing them in the database;

[0038] S500, the server compares the vibration displacement data, vibration frequency data, temperature and humidity data of the target device, maximum temperature, minimum temperature and average temperature data with the threshold values ​​set in advance in the database respectively. When at least one of the data exceeds the threshold value set in advance in the database for more than a specified specific time period, the server determines that the target detection device is in an abnormal working state and generates a warning message;

[0039] S600, the server sends the generated warning information and the maximum temperature, minimum temperature and average temperature data of the current device, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection device to the mobile terminal or web page for graphical display and editing.

[0040] The thermal infrared camera is a thermal imaging device with a Type-c interface, which can be connected to a mobile terminal and can be equipped with a thermal imaging algorithm model to generate corresponding thermal infrared images. The thermal imaging algorithm of the present invention performs noise filtering and temperature region segmentation on the image and performs pixel-level temperature calculation on the image, thereby improving the clarity of the image and the accuracy of the thermal distribution map. The implementation steps are as follows: Figure 2 As shown:

[0041] S201, noise filtering and temperature region segmentation are performed on the thermal infrared image. Although the traditional Gaussian filter can effectively remove noise, in the complex environment of the offshore platform, the lighting conditions are unstable, resulting in various types of noise in the image. The traditional Gaussian filter using fixed weights cannot effectively remove noise from the image. Therefore, the present invention proposes an improved Gaussian filtering method to perform noise filtering and temperature region segmentation on the thermal infrared image, improve the image analysis accuracy of the offshore platform, and eliminate errors caused by environmental or equipment noise. The improved Gaussian filter introduces a weight w(i, j) parameter to adjust the influence of different pixels in the filtering process. Each pixel value I′(x, y) after filtering can be calculated by the following formula:

[0042]

[0043] Among them, I(x,y) is the pixel value of the original image, G(i,j) is the two-dimensional Gaussian kernel function, k is the size of the filter window, W is the normalization factor, which ensures that the pixel value after filtering remains within a reasonable range, and w(i,j) is the weight function, which is dynamically adjusted according to the noise characteristics of local pixels. The filter can more accurately identify and remove noise and adapt to the complex situation of offshore platforms, thereby improving the accuracy of image analysis.

[0044] S202, after removing the image noise, the present invention further proposes a method based on pixel-level temperature calculation to identify the hot spot area on the surface of the equipment. According to the principle of thermal imaging, a thermal distribution map is generated by converting the infrared radiation intensity into temperature. In order to better adapt to the special environment of the offshore platform, compensate for the influence of environmental factors (such as humidity, wind speed, etc.) on the measurement, and ensure the accuracy of the temperature reading, the present invention has revised the traditional temperature conversion formula and added a correction factor C. The specific formula is as follows:

[0045]

[0046] Where L is the infrared radiation intensity and σ is the Stefan-Boltzmann constant 5.67×10 -8 W / m 2 K 4 , T is the absolute temperature in Kelvin, ∈ is the emissivity of the device surface, and the emissivity of the device surface is 0.95; C is the correction factor, which is 0.9 at the upper temperature limit and 1.1 at the lower temperature limit to ensure that the temperature of thermal infrared recognition is within this range; by calculating each pixel point in the image, the radiation intensity of each pixel is converted into the corresponding temperature value.

[0047] The generated thermal infrared image is sent to the server for temperature extraction. The present invention provides a text recognition method for identifying temperature data from a thermal infrared image. The method includes three steps: contrast enhancement, text area segmentation, and text recognition. The method can effectively identify the temperature value in the thermal infrared image. The specific steps of the text recognition algorithm are as follows: Figure 3 As shown:

[0048] S301, firstly, contrast enhancement and image sharpening processing are performed on the text area in the thermal infrared image. Considering the grayscale characteristics of the thermal infrared image, the traditional contrast enhancement method may not be able to fully highlight the text area. Therefore, the present invention optimizes the traditional contrast adaptive histogram equalization (CLAHE) algorithm, and achieves a better contrast enhancement effect by adjusting the cumulative distribution function (CDF), ensuring that the text area can be better enhanced when processing the thermal infrared image. The CLAHE algorithm formula is as follows:

[0049]

[0050] Among them, CDF(I(x,y)) is the cumulative distribution function of the pixel value I(x,y), M×N is the total number of pixels in the image, L is the gray level, CDF min is the minimum non-zero value in the cumulative distribution function. This algorithm is used to enhance the text part in the image to make it clearer.

[0051] S302, after contrast enhancement and image sharpening processing are performed on the text area in the thermal infrared image, the text area is identified. The thermal infrared image of the offshore platform will contain a large amount of non-text information, such as equipment outlines, background noise, etc., which interferes with the recognition of text. The DBNet text detection model can be used to identify the area containing text in the image. The DBNet model uses deep learning technology to effectively segment the text part in the thermal infrared image by generating a binary probability map, thereby improving the accuracy of text area detection;

[0052] S303, using the text recognition model to recognize the text in the text area identified by DBNet. The text in the thermal infrared image is small and has low contrast. The present invention adopts a convolutional recurrent neural network (CRNN) structure, extracts image features through multi-layer convolution, and introduces a bidirectional LSTM structure with an attention mechanism, thereby improving the ability of CRNN in processing sequence text, so that the network can better capture and recognize small-scale, low-contrast text in thermal infrared images. The bidirectional LSTM formula with an attention mechanism is as follows:

[0053] h t =Attention(x t ,h t-1 )·σ(W h ·x t +U h ·h t-1 +b h )

[0054] Among them, x t is the input feature vector, h t is the hidden state at the current moment, W h and U h is the weight matrix, b h is the bias term, Attention(x t ,h t-1 ) is an attention mechanism that calculates the importance of each character based on the current input and historical hidden states. CRNN uses this model to achieve sequence-to-sequence conversion and ultimately outputs the recognized maximum temperature, minimum temperature, and average temperature data.

[0055] The highest, lowest and average temperatures extracted by the text recognition algorithm and the vibration displacement data, vibration frequency data, temperature data and humidity data collected by the vibration sensor and temperature and humidity sensor installed on the equipment are uploaded to the server and stored in the database. When at least one data exceeds the corresponding data threshold set in advance for more than a specified specific time length, the server will determine that the equipment is in an abnormal working state, give a warning message, and feed back the result to the mobile terminal. The warning information includes the corresponding data information of the highest temperature, lowest temperature and average temperature data, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection equipment that exceeds the threshold set in advance in the database.

[0056] The server extracts the vibration displacement data, vibration frequency data, temperature data, humidity data, maximum temperature, minimum temperature and average temperature from the database, and displays and edits them graphically on the web page. The graphical display presents the current operating status and data change trend of the equipment in the form of data curves and data tables, including line graphs and bar graphs, so that users can monitor the changes of equipment operating parameters over time in real time.

Claims

1. A thermal infrared imaging detection and early warning method for offshore platform equipment based on a mobile terminal, which is suitable for real-time monitoring and early warning of offshore platform target detection equipment, wherein the offshore platform target detection equipment is equipped with a vibration sensor and a temperature and humidity sensor, and the method comprises the following steps: S100, the mobile terminal is connected to a thermal infrared camera, and the mobile terminal receives the location and name of the offshore platform target detection device input by a user; S200, photographing an image of a target detection device by means of the thermal infrared camera, extracting a temperature distribution area of ​​the target detection device by means of a thermal imaging algorithm, extracting temperature data from the thermal infrared image, and uploading the data to a server; S300, the server uses a text recognition algorithm to recognize the thermal infrared image, extracts the maximum temperature, minimum temperature and average temperature data in the thermal infrared image, and stores the maximum temperature, minimum temperature and average temperature data in a database; S400, obtaining vibration displacement data and vibration frequency data of the target device through the vibration sensor installed on the target detection device, obtaining temperature and humidity data of the target detection device through the temperature and humidity sensor installed on the target detection device, and uploading the collected thermal infrared image, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection device to the server, and storing them in the database; S500, the server compares the vibration displacement data, vibration frequency data, temperature and humidity data of the target device, maximum temperature, minimum temperature and average temperature data with the threshold values ​​set in advance in the database respectively. When at least one of the data exceeds the threshold value set in advance in the database for more than a specified specific time period, the server determines that the target detection device is in an abnormal working state and generates warning information of the target detection device; S600, the server sends the generated warning information to a mobile terminal or a web page for graphical display and editing.

2. The method according to claim 1, characterized in that The offshore platform target detection equipment is one or more of a gas compressor, a heat exchanger, an explosion-proof motor, a gas storage tank, a gas flotation machine, an electric dehydrator, and an inclined plate deoiler.

3. The method according to claim 1, characterized in that The thermal infrared camera is a thermal imaging device with a Type-c interface, which is connected to a mobile terminal via the Type-c interface and is equipped with a thermal imaging algorithm to generate corresponding thermal infrared images.

4. The method according to claim 1, characterized in that The thermal imaging processing algorithm includes: S201, performing noise filtering and temperature region segmentation on the thermal infrared image. In order to improve the image analysis accuracy of the offshore platform and eliminate errors caused by environmental or equipment noise, an improved Gaussian filter is used; the improved Gaussian filter introduces a weight w(i, j) parameter to adjust the influence of different pixels in the filtering process, and each pixel value I′(x, y) after filtering is calculated by the following formula; Among them, I(x,y) is the pixel value of the original image, G(i,j) is the two-dimensional Gaussian kernel function, k is the size of the filter window, W is the normalization factor to ensure that the pixel value after filtering remains within a reasonable range, and w(i,j) is the weight function, which is dynamically adjusted according to the noise characteristics of the local pixel to adapt to the complex situation of the offshore platform; S202, through image pixel-level temperature calculation, identify the hot spot area on the surface of the equipment, generate the corresponding thermal distribution map, and convert the infrared radiation intensity into temperature based on the principle of thermal imaging and the particularity of the offshore platform environment. The specific formula is as follows: Where L is the infrared radiation intensity and σ is the Stefan-Boltzmann constant 5.67×10 -8 W / m 2 K 4 , T is the absolute temperature in Kelvin, ε is the emissivity of the device surface, and the emissivity of the device surface is 0.95; C is the correction factor, which is 0.9 at the upper temperature limit and 1.1 at the lower temperature limit to ensure that the temperature of thermal infrared recognition is within this range; by calculating each pixel point in the image, the radiation intensity of each pixel is converted into the corresponding temperature value.

5. The method according to claim 1, characterized in that The text recognition method comprises: S301, before performing text recognition, firstly perform contrast enhancement and image sharpening processing on the text area in the thermal infrared image, and improve the text clarity in the thermal infrared image and increase the recognizability of the text by using the contrast adaptive histogram equalization (CLAHE) algorithm. The CLAHE algorithm formula is as follows: Among them, CDF(I(x,y)) is the cumulative distribution function of the pixel value I(x,y), M×N is the total number of pixels in the image, L is the gray level, and CDF min is the minimum non-zero value in the cumulative distribution function. The text part in the image is enhanced by this algorithm to make it clearer; S302, identifying the area containing text in the thermal infrared image through the DBNet text detection model, effectively segmenting the text part in the thermal infrared image, and using the DBNet algorithm for text area detection to generate the text area through the binarization probability map; S303: For the detected text area, a text recognition model is used for recognition. The model adopts a convolutional recurrent neural network (CRNN) structure, extracts image features through multi-layer convolution, and uses a bidirectional LSTM with an attention mechanism to perform character-level sequence decoding, so that the network can better capture the dependency between contexts and finally output the text content. The formula of the bidirectional LSTM with an attention mechanism is as follows: h t =Attention(x t ,h t-1 )·σ(W h ·x t +U h ·h t-1 +b h ) Among them, x t is the input feature vector, h t is the hidden state at the current moment, W h and U h is the weight matrix, b h is the bias term, Attention(x t ,h t-1 ) is an attention mechanism that calculates the importance of each character based on the current input and historical hidden states. CRNN realizes sequence-to-sequence conversion through this model and finally outputs the highest temperature, lowest temperature and average temperature data in the recognized thermal infrared image.

6. The method according to claim 1, characterized in that The database adopts MySQL database, which stores the vibration displacement data, vibration frequency data, temperature and humidity data of the target device, maximum temperature, minimum temperature and average temperature data, fault records, and early warning information of the device, and performs correlation analysis on the above information through data mining technology to improve the efficiency of equipment management and maintenance.

7. The method according to claim 1, characterized in that The warning information includes corresponding data information of the highest temperature, lowest temperature and average temperature data, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection device that exceed the threshold value set in advance in the database.

8. The method according to claim 1, characterized in that: The server extracts the data table information from the database and displays it on the web page. Users access the web page through a browser to view the maximum temperature, minimum temperature, average temperature data, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection equipment in real time, and edit and manage the data. Users view this information in real time through the web page and analyze historical warning information.

9. The method according to claim 1, characterized in that: The warning information is transmitted to the mobile terminal in real time via Socket communication, and the device location and detailed warning information of data exceeding the threshold are displayed on the application interface of the mobile terminal.

10. The method according to claim 1, characterized in that The warning information includes corresponding data information of the highest temperature, lowest temperature and average temperature data, vibration displacement data, vibration frequency data, temperature and humidity data of the target detection device that exceed the threshold value set in advance in the database.

Images