CN115825075A - Cable breakage detection method based on infrared signal - Google Patents

Abstract

The method and system for detecting cable breakage based on infrared signals relate to the technical field of cable fault detection, and can directly observe the heat distribution state of an energized cable, and obtain an infrared thermal image of the cable by using a shooting function. Infrared thermal images have all the characteristics of general images, and most of the information in the images is contained in the edge features. Significant changes in these characteristic attributes usually reflect important events and changes. These include discontinuities in depth, discontinuities in surface orientation, changes in material properties, and changes in scene lighting. This paper improves the Otsu segmentation method, which can find the fault point in the infrared image of the damaged cable, so that the system can diagnose the damaged part of the cable more intelligently. The system connects the control module, the image acquisition module, the image processing module and the alarm module together, and uses Python programming to complete the system.

Description

Cable breakage detection method based on infrared signal

technical field

The invention relates to the technical field of cable fault detection, in particular to a method for detecting local defects of cables based on infrared signals.

Background technique

There may be insulation defects in the cable body due to various reasons. The increase in resistance at the cable insulation defect will increase the loss and increase the temperature, further deteriorating the insulation state. In the process of cable temperature rise, there are usually three ways to transfer heat energy: heat conduction, heat convection and heat radiation. The heat conduction is caused by the temperature gradient between the various parts in the cable, and the heat convection and heat radiation are caused by the temperature difference between the cable and the outside air.

From the perspective of infrared diagnosis, faults can be roughly divided into two categories: external faults and internal faults. External faults refer to the poor operating state of the cables exposed in the natural environment (poor contact at the joints, dirty cable surfaces, or degradation of external insulation performance due to mechanical effects, etc.). Such faults occur outside, and the thermal radiation energy field formed by the heat generated by them It can be well captured by the infrared thermal imager, thus forming an infrared diagnostic map with obvious light and dark distribution; internal fault refers to the abnormal temperature rise phenomenon that occurs on the conductive and insulating medium inside the cable (damp and aging cause the loss of the insulating medium to increase). Large defects, increased ferromagnetic loss caused by internal defects, uneven voltage distribution and increased leakage current defects, poor conductor connection or increased contact resistance loss, etc.), such faults are usually not covered by the internal insulation medium and equipment casing Obtain intuitive fault status information like external faults. Combined with the knowledge of heat transfer, through the infrared image analysis of a large number of on-site inspections and experimental studies, the type, location and severity of internal faults or defects of the equipment can also be detected according to the temperature distribution law on the outer surface of the equipment.

Through the analysis of the heating characteristics of the cable and the comprehensive consideration of the portability of the device, the infrared detection device designed in the present invention needs to follow the characteristics of easy portability, clear imaging and simple system.

The infrared detection system of the present invention is a non-contact thermal imaging instrument based on an infrared array high-precision temperature sensor and an advanced software algorithm, which can perform infrared imaging on any object within the field of view, with an imaging resolution of 512*384 pixels and a temperature sensitivity of 0.1 ℃, the absolute accuracy is ±1.5℃, and the refresh rate is up to 64Hz. Built-in storage and real-time clock, with real-time data output display, photo storage function, very convenient to use.

Contents of the invention

The cable breakage detection method based on infrared signal is characterized in that:

1), the construction of the system operating environment:

Install the operating system used by the Raspberry Pi, set a static IP address for the Raspberry Pi, install the OpenCV vision library, and develop software on this basis;

2), development and debugging of infrared detection system:

Raspberry Pi 4B combined with MLX90640 infrared sensor to build an infrared detection module; the resolution of MLX90640 is 32*24 pixels, using cubic polynomial combined with bilinear interpolation, after two interpolations, the image is expanded from 32*24 pixels to 512* 384 pixels;

With the Raspberry Pi as the main control, supplemented by an infrared imaging module, the infrared video can be displayed in real time on the computer through a remote wireless connection, or can be carried around to form a portable infrared detection device;

3), infrared detection experiment:

When the cable is running, the temperature at the fault point rises significantly, the temperature at both ends of the fault is higher than that in the middle of the fault, and the infrared sensor detects the fault point;

4), infrared image segmentation algorithm

The threshold segmentation method is used to extract the grayscale data of the target layer and the background layer in the image, and the grayscale difference is obtained to realize image separation; the optimized image segmentation method is as follows:

According to the following formula (1), the maximum variance threshold is obtained:

S＝w ₀ *(uu ₀ ) ² +w ₁ *(uu ₁ ) ² (1)

w ₀ refers to the proportion of the background point in the image; u ₀ refers to the average gray value of the background point; w ₁ refers to the average gray value of the foreground point; u refers to the average gray value of the image degree; S refers to the image variance, that is, the maximum variance threshold used;

5), infrared image edge detection

In order to determine the defect severity of the defective part of the cable, the suspected overheated area of the cable is extracted; the RGB component of the pixel corresponding to the overheated part is higher than that of the normal part; the K-Means clustering algorithm is used for classification, and the k value ranges from 4 to 7.

The example verification shows that the continuity, accuracy and noise resistance of the infrared image detection results obtained by using this algorithm are relatively good, and the edge information such as the abnormal heating position and range of the infrared image in a noisy environment can be effectively extracted, and then according to The difference of the abnormal hot edge of the infrared image is extracted under different operating states, which is suitable for the identification and classification of the abnormal heating state.

Description of drawings

Figure 1 System software operating environment construction block diagram

Figure 2 Effect diagram of polynomial interpolation transformation

Figure 3 fault wire image

Figure 4 Infrared image of the faulty wire

Figure 5 Infrared image processing results of faulty wires

Figure 6 Connection failure image

Figure 7 fault image of the middle part

Figure 8 kmeans algorithm processing intermediate fault infrared image

Figure 9 kmeans algorithm processing connection fault infrared image

Detailed ways

1. Construction of the system operating environment:

The Raspberry Pi itself does not contain an operating system. It needs to be installed by itself, and then configure the operating environment of the program, and develop applications on this basis. Figure 1 shows the steps needed to build the infrared detection system based on the Raspberry Pi. The operation environment of raspberry group can install the operating system that uses to comprise windows, Android, IOS, Linux etc., the present invention selects Linux operating system for use; Static IP address is set for raspberry group, convenient debugging; OpenCV visual library is installed, on this basis Do software development.

2. Development and debugging of infrared detection system:

The present invention uses Raspberry Pi 4B combined with MLX90640 infrared sensor to build an infrared detection module. The resolution of MLX90640 is 32*24 pixels, and the image is expanded from 32*24 pixels to 512*384 pixels after two interpolations by using cubic polynomial combined with bilinear interpolation. The effect is shown in Figure 2. First cubic polynomial interpolation and then bilinear interpolation, the high and low temperature distribution is more obvious.

With the Raspberry Pi as the main control, supplemented by a miniature infrared imaging module, the infrared video can be displayed in real time on the computer through a remote wireless connection, and it can also be carried around to form a portable infrared detection device, which is convenient for inspection.

3. Infrared detection experiment:

When the cable is running, the temperature at the fault point rises obviously, and the temperature at both ends of the fault is higher than that in the middle of the fault. The infrared sensor can accurately detect the fault point. Figure 3 is the original infrared image of the faulty wire, and Figure 4 is the infrared image of the faulty wire.

4. Infrared image segmentation algorithm

The invention adopts the threshold segmentation method to extract the grayscale data of the target layer and the background layer in the image, obtains the grayscale difference, and realizes image separation. The present invention adopts the optimized image segmentation method as follows:

According to the following formula (1), the maximum variance threshold is obtained:

S＝w ₀ *(uu ₀ ) ² +w ₁ *(uu ₁ ) ² (1)

w ₀ refers to the proportion of the background point in the image; u ₀ refers to the average gray value of the background point; w ₁ refers to the average gray value of the foreground point; u refers to the average gray value of the image degree; S refers to the image variance, which is the threshold used in the present invention.

As shown in Figure 5, the left image is the original infrared image of the faulty wire, and the right image is the image after threshold segmentation.

5. Infrared image edge detection

In order to determine the defect severity of the defective part of the cable, it is necessary to extract the suspected overheating area of the cable. The RGB components of pixels corresponding to overheated parts are higher than normal parts. The present invention adopts K-Means clustering algorithm to classify:

Experiments show that the choice of k value will affect the final result of clustering. Because the cables have different backgrounds in the actual cable arrangement, and the severity of cable faults is different, the present invention selects the value of k to be in the range of 4-7.

According to the state detection experiment of cable insulation fault, two infrared original images are obtained. Figure 6 is the original image of a fault at the connection of two cables, and Figure 7 is the original image of a fault in the middle of the cable, and the fault point is marked with a rectangular frame in the figure. Using the K-Means clustering algorithm to process these two infrared images respectively, as shown in Figure 8 and Figure 9, the fault point can be clearly observed. In FIGS. 8 and 9 , the left image is the segmented image, and the right image is the clustered image.

Any obvious changes made by those skilled in the art based on the teachings of the present invention belong to the protection scope of the present invention.

Claims (1)

1. The cable breakage detection method based on infrared signal, is characterized in that: 1), the construction of the system operating environment: Install the operating system used by the Raspberry Pi, set a static IP address for the Raspberry Pi, install the OpenCV vision library, and develop software on this basis; 2), development and debugging of infrared detection system: Raspberry Pi 4B combined with MLX90640 infrared sensor to build an infrared detection module; the resolution of MLX90640 is 32*24 pixels, using cubic polynomial combined with bilinear interpolation, after two interpolations, the image is expanded from 32*24 pixels to 512* 384 pixels; With the Raspberry Pi as the main control, supplemented by an infrared imaging module, the infrared video can be displayed in real time on the computer through a remote wireless connection, or can be carried around to form a portable infrared detection device; 3), infrared detection experiment: When the cable is running, the temperature at the fault point rises significantly, the temperature at both ends of the fault is higher than that in the middle of the fault, and the infrared sensor detects the fault point; 4), infrared image segmentation algorithm The threshold segmentation method is used to extract the grayscale data of the target layer and the background layer in the image, and the grayscale difference is obtained to realize image separation; the optimized image segmentation method is as follows: According to the following formula (1), the maximum variance threshold is obtained: S＝w ₀ *(uu ₀ ) ² +w ₁ *(uu ₁ ) ² (1) w ₀ refers to the proportion of the background point in the image; u ₀ refers to the average gray value of the background point; w ₁ refers to the average gray value of the foreground point; u refers to the average gray value of the image degree; S refers to the image variance, that is, the maximum variance threshold used; 5), infrared image edge detection In order to determine the defect severity of the defective part of the cable, the suspected overheated area of the cable is extracted; the RGB component of the pixel corresponding to the overheated part is higher than that of the normal part; the K-Means clustering algorithm is used for classification, and the k value ranges from 4 to 7.

Images