CN102565625A - Method for intelligently diagnosing thermal defects of high-voltage transmission line based on infrared image - Google Patents

Abstract

The invention relates to a method for intelligently diagnosing thermal defects of high-voltage transmission lines based on infrared images. It is a method for automatically detecting thermal defects in transmission lines from captured infrared images in a moving state. The operation steps are as follows: Step 1. Use line object perception clustering to identify components in the line; effectively eliminate the influence of heat generated by cars driving on the ground in natural and complex scenes, heat reflected by ground heating objects and poles and towers on the diagnosis of line thermal defects. Step 2. For the identified line components, use the relative temperature difference method to diagnose whether the component has thermal defects, which can eliminate the impact of additional temperature rise caused by solar radiation. At the same time, the parameters such as detection distance, ambient temperature, humidity, and wind speed are not accurate. Errors are also reduced. The method can effectively improve the efficiency of thermal defect detection of transmission lines, and can be effectively applied to vehicle-mounted or helicopter transmission line inspection services.

Description

Method of Intelligent Diagnosis of Thermal Defects in High Voltage Transmission Lines Based on Infrared Images

technical field

The invention relates to a method for intelligently detecting thermal defects of high-voltage transmission lines based on infrared thermal images. In particular, it relates to a method for intelligently detecting thermal defects existing in transmission lines from captured infrared images in a moving state, and the method can effectively improve the efficiency of detecting thermal defects of transmission lines. This method can be effectively applied to vehicle or helicopter transmission line inspection business.

Background technique

The high-voltage transmission line is the artery of the power system. Its operating status directly determines the safety of the power system and the operation of the national economy. Infrared detection has the characteristics of long-distance, no power failure, no contact, and no disassembly. Thermal defects provide an advanced means for power system line condition monitoring.

Infrared thermal imaging camera uses infrared detector, optical imaging objective lens and focal plane light sensing system to accept the infrared radiation energy distribution pattern of the measured target and reflect it on the photosensitive element of the infrared detector. Between the optical system and the infrared detector, the focal plane The planar light sensing system focuses the infrared thermal image of the measured object on the infrared detector, and the infrared radiation energy is converted into an electrical signal by the detector, and the infrared thermal image is displayed on the TV screen or monitor after amplification, conversion or video signal .

There are many factors that affect the measurement accuracy of infrared thermal imaging cameras outdoors. For example, Gao Xiaoming, Analysis of Factors Affecting Measurement Accuracy of Infrared Thermal Imager, Huadian Technology, November 2008, Volume 30, Issue 11, analyzed these factors: different measured objects have different emissivity. The emissivity of an object is between 0 and 1, and its size has a certain relationship with the material, surface roughness, shape, oxidation degree, color, thickness, etc. of the object. For metal materials of transmission lines, the surface state has a great influence on the emissivity. Generally, the emissivity of the rough surface and the oxidized surface is several times that of the polished surface; the emissivity is related to the angle of measurement, and the larger the angle of measurement , the greater the error; the environment has a greater impact on the infrared temperature measurement work: in the transmission process of infrared radiation, due to the absorption of the atmosphere, there will always be a certain energy attenuation, in the atmosphere close to the ground, the absorption of infrared radiation energy The main gases are water vapor and carbon dioxide. At present, most infrared thermal imaging cameras do not have compensation methods for atmospheric attenuation; when sunlight or strong light is irradiated, due to light reflection and diffuse reflection, it will greatly affect the normal operation and accurate judgment of infrared thermal imaging cameras. At the same time, The temperature rise of the object under test caused by light irradiation will be slightly added to the stable temperature rise of the device under test, which will directly cause measurement errors. Therefore, it is best to choose a cloudy day without sunlight for infrared temperature measurement; When it is windy, the flow of air will accelerate the heat dissipation on the surface of the heating object, reducing the surface temperature of the object; the dust and suspended particles in the atmosphere are another reason for the energy attenuation of infrared radiation during the propagation process; the influence of thermal radiation of adjacent objects on temperature measurement , the lower the temperature of the measured object or the smaller the emissivity, the greater the impact of thermal radiation from adjacent objects, and the greater the impact on the temperature measurement accuracy; when a component in the series circuit of the transmission line has an overheating defect, the running The electrical connector will conduct heat to the surrounding components, causing other components in the circuit near the hot spot to heat up. However, due to the loss in the heat transfer process, the temperature of these points is lower than the hot spot. Therefore, when measuring the temperature, it must be done To accurately locate, find the real heat source.

Infrared thermal imaging cameras are often used to detect transmission line defects, such as Cheng Yulan, Field Practical Technology of Infrared Diagnosis [M ]. Beijing: Mechanical Industry Press, 2002, 4. Similar comparison method: compare the temperature rise of the same type of components at different positions on the same line, and then determine whether there is a thermal defect; historical comparison method: perform temperature rise on the thermal images of the same type of components at the same position taken at different times on the same line Compare, and then judge whether there is a thermal defect. The two thermal defect analysis methods, the similar comparison method and the historical comparison method, are relatively reliable, but the efficiency is low, and a large amount of measurement data is required for analysis. Absolute temperature alarm limit temperature rise method: A temperature threshold is specified according to the wire type and load current. If the temperature exceeds this threshold, there may be thermal defects. Since the environment has a great influence on the measured object, such as the temperature rise caused by solar radiation, this method The method is neither reliable nor accurate.

Hu Shizheng, Relative Temperature Difference Judgment Method and Judgment Standard for Infrared Diagnosis of Electrical Equipment, Power Grid Technology, October 1998, Volume 22, Issue 10. The relative temperature difference refers to the temperature difference between two corresponding measuring points with the same or basically the same condition (referring to the equipment model, installation location, ambient temperature, surface condition and load size), and the percentage of the temperature rise ratio of the hotter point.

Guo Xianxiao, Li Wei, Cai Hansheng, A Preliminary Study on Infrared Detection of High Voltage Transmission Lines, HIGH VOLTAGE ENGIINEERING, June.1999, Vol.25, No.2. In the infrared thermal defect detection of transmission lines, the absolute temperature difference method is often used: take the highest temperature of the normal operating wire or line fittings 1m away from the measured object as the reference temperature T _a , and the temperature of the measured object is T , △T = T - T _a , judge the thermal defect according to △T , this method can eliminate the influence of additional temperature rise caused by solar radiation. At the same time, the error caused by the inaccuracy of parameters such as isotropy, detection distance, ambient temperature, humidity, and wind speed is also reduced. At full load, ΔT exceeds 5 for judging the heat generation of the high-voltage line ^. At C, it can be considered as a slight contact hazard (general thermal defect), and △T exceeds 15 ^. C is a major defect, and △T exceeds 40 ^. C is urgent defect.

Yu Deming, Shen Jian, Wang Jun, Yao Wenjun, Chen Fangdong, Wu Yi, Comparative Analysis of Helicopter and Manual Inspection Effects, China Electric Power, November 2008, Volume 41, Issue 11. At present, both at home and abroad are starting to study the inspection system based on helicopters and robots, analyze and compare the advantages and disadvantages of helicopter inspection and manual inspection, and point out that helicopter inspection has incomparable advantages in the ability to inspect hidden defects of equipment. In the helicopter inspection technology, due to the influence of aviation control, the development of our country lags behind, and the research and development really started in the early 21st century. The equipment for detecting defective components is generally developing towards the direction of multi-spectral image synthesis, and the diagnosis and detection technology of high-voltage line components has been developing towards the direction of intelligence.

When helicopters are used to conduct inspections, thermal imaging cameras are an important means of detecting thermal defects in lines. It is hoped that the use of thermal imaging cameras can automatically find thermal defects in the line. Due to the interference of the ground background, such as the influence of vehicles driving on the ground, ground heating elements and the reflection heat of the tower itself, it is very difficult to automatically find the thermal defects of the line. The invention adopts the combination technology of infrared image recognition and relative temperature difference method. Intelligent identification of circuit components on the infrared image, on the basis of the identification of the components, to find the highest temperature, and then use the relative temperature difference method to determine whether there are thermally defective components. This method can be effectively applied to the intelligent detection of thermal defects in high-voltage lines during helicopter or vehicle inspections.

Contents of the invention

In view of this, the purpose of the present invention is to intelligently diagnose the thermal defects of the line through the infrared image collected by the infrared thermal imager during the helicopter inspection or the vehicle inspection, that is, to provide an intelligent high-voltage transmission line thermal defect based on the infrared image. The defect method can effectively improve the efficiency of thermal defect diagnosis when inspecting high-voltage lines.

In order to achieve the above object, the design of the present invention is as follows: the present invention requires that the background interference can be effectively eliminated during the helicopter inspection or vehicle inspection, such as the heat source sent by the automobile on the ground in the natural complex background or the ground heating body, the pole tower itself in the sun. The heat generation under long-term irradiation, these heat source objects and the heating of circuit components are superimposed on the infrared image, which often affects the judgment of the actual temperature of circuit components, resulting in a large number of misjudgments. On transmission lines, infrared heat defects often appear in wires, drain wires, insulators and the contact parts of these components. Therefore, only by identifying the conductors, drain wires, and insulator components in the transmission line can the thermal defects of the line be automatically and reliably diagnosed.

The invention combines infrared image line component identification with relative temperature difference judgment technology to intelligently realize the diagnosis of thermal defects of line components. In the identified temperature area of each component, automatically find the highest temperature, use the highest temperature pixel as the seed point, use the region growing method, and use 5 degrees below the highest temperature as the edge judgment condition to generate a highest temperature area, such as the highest temperature If the number of temperature regions is approximately the same as the number of temperature regions of the part, the temperature of the part is normal; if the number of the highest temperature region is obviously smaller than the number of temperature regions of the part, the pixels outside the highest temperature region of the part are regarded as a region to calculate the maximum temperature The average temperature in the area; calculate the average temperature of the area except the highest temperature, and use the relative temperature difference method to diagnose whether there is a thermal defect in the component. The relative temperature difference method can eliminate the influence of additional temperature rise caused by solar radiation, and at the same time, reduce the errors caused by inaccurate parameters such as detection distance, ambient temperature, humidity, and wind speed.

From the perspective of the transmission line scene taken by the infrared thermal imager on the helicopter: it is composed of a natural background and a transmission line. From the perspective of the composition and structure of the transmission line: it is composed of a straight pole tower, a tension pole tower, and a wire and a ground wire between the pole towers.

Due to some heating objects in the natural scenery, driving cars, and towers themselves heating under the sunlight, these heating objects often interfere with the diagnosis of thermal defects of the line by the infrared thermal imager, causing misjudgment and directly affecting the accuracy of the diagnosis of infrared thermal defects and reliability. If professional infrared analysis software is used to analyze the thermal defect of the line on the infrared image, a large number of interactive operations are required. First, the human eye is used to find out the line equipment from the infrared image, and the regional maximum temperature analysis method is used to find out whether there is a thermal defect. During helicopter inspections, a large amount of infrared image data is used in real time, and interactive analysis and diagnosis are used. The work intensity is high, and the efficiency of thermal defect diagnosis is low.

In order to improve the efficiency of infrared thermal defect diagnosis, during helicopter inspection or vehicle inspection, it is possible to collect infrared images while diagnosing thermal defects on the line, thereby improving the efficiency of infrared thermal defect diagnosis. Therefore, it is necessary to identify circuit components with current loads from the infrared image, such as wires, drain wires, insulators, and the contact areas of these components, and then use the relative temperature difference method to diagnose the thermal defects of the components on each identified component area.

To identify circuit components on infrared images, first establish a knowledge model of the circuit and the positional relationship between the upper and lower components. The high-voltage line can be regarded as composed of line segments in different directions: the conductor is composed of multiple parallel line objects, the drain line is a downward curved curve, and the tower is composed of line segments in different directions. The high voltage has a stable relational connection with a fixed position.

Classify the collected infrared images. Since the thermal imaging camera cannot capture all the 500-meter line corridors between the two towers, the collected infrared thermal images of the lines are divided into three categories: the first type of infrared images has no towers, only wires, and how many wires are there. It is composed of parallel straight line objects and runs through the whole infrared image; the second type of infrared image has both wires and towers and is a straight line; the third type of infrared image is defined as: there are both wires and towers in the image and For the tension tower.

According to the above inventive concept, the present invention adopts the following technical solution: a method for intelligently diagnosing thermal defects of high-voltage lines based on infrared images, which is characterized in that the operation steps are as follows:

1. Use the line object perception clustering algorithm to identify line components. The specific steps are as follows:

1-1. The edge operator with direction processes the infrared image, extracts horizontal line segments, vertical line segments, oblique line segments, and curved line segments, and generates infrared binary images through the maximum inter-class double difference method, which is denoted as IB(i,j) Infrared binary image;

1-2. On the infrared binary image in step 1-1, calculate the slope and intercept of each horizontal, inclined, and vertical small line segment, and merge and connect the small line segments with the same slope and intercept into long straight line segments; Fit curve segments by analyzing the common endpoints of small line segments;

1-3. Divide the area vertically on the infrared binary image in step 1-2, and analyze the distribution density of three types of small line segments in each area; if there is an area with high distribution density of horizontal, vertical, and inclined line segments at the same time, confirm it as Contains the infrared image of the tower; if there is no curve in the image of the tower, it is confirmed to contain the infrared image of the straight tower; if there is a curve in the image of the tower, it is confirmed to contain the infrared image of the tension tower;

1-4. Among the horizontal line segment and the inclined line segment, the line segment with approximately the same slope and different intercept is classified as a parallel line group; and confirmed as a wire group;

1-5. According to steps 1-3 and 1-4, if there are two sets of parallel wires in the infrared image that contains straight poles and towers, judge the proximity of adjacent endpoints, and at the same time expand the wires relatively so that the two sets of wires intersect , there is a synthetic insulator in the vertical direction of the intersection point, and the position of the synthetic insulator is deduced as the identified synthetic insulator part; if there is a group of parallel wires, the endpoints adjacent to the tower area are expanded according to the 16*8 pixel area, so that It effectively extends to the contact area between line components;

1-6. According to steps 1-3 and 1-4, if there is a group of parallel wires in the infrared image that includes the tension tower, determine the end point of the wire adjacent to the curve, and there is a glass insulator component in the parallel extension direction of the end point , deduce the position of the glass insulator as the identified glass insulator part;

1-7. Manage the position coordinates of the identified conductors, drain wires, and insulators with an independent connected area structure, specifically adopt the following structure management for the management of identified components;

struct part_object {

PART_NAME part_ID;

unsigned int elements_number;

        vector<struct Element> elements;

                                           //The coordinates of the rectangular area bounded by the component

int leftx, lefty;

int topx, topy;

int rightx, righty;

int bottomx, bottomy;

}

On the infrared image, the location coordinates of each component area will be identified and managed with the object area BLOB, and the mapping between the pixels of each component area and the corresponding actual temperature will be established, and then the temperature area of each component will be established.

2. In the identified area of each component, use the relative temperature difference method to diagnose its thermal defects. The specific thermal defect diagnosis adopts the following process:

2-1. Read the identified area of each component, and read the actual temperature value in the area from the infrared temperature image IW(i, j) according to the pixel coordinates of the area, and automatically find out the wire parts in blocks The maximum temperature value, according to the connectivity of the insulator parts, automatically find the maximum temperature value, recorded as T _{max_hot} ;

2-2. Take the highest temperature pixel as the seed point, use the region growing method, and use 5 degrees below the highest temperature as the boundary judgment condition to generate connected regions with the highest temperature. For example, the size of the highest temperature region is similar to the size of the component region If it is consistent, the temperature of the component is normal, and the diagnosis is ended; if the number of the highest temperature area is obviously smaller than the size of the area of the component, the pixels outside the highest temperature area of the component are regarded as an area, and then go to step 2-3;

2-3. Generate two temperature connected regions on the infrared temperature image IW(i,j) for this component region: one is the connected region with the highest temperature, and the other is the connected temperature region composed of pixels other than the highest temperature. Calculate the average value AVG_T _max of the temperature in the connected region with the highest temperature, and calculate the average value AVG_T _ref of the temperature in the connected region composed of pixels other than the highest temperature, as follows:

诊断该部件热缺陷： 2-4. Basis

Diagnose the component thermal defect:

 ，则该部件有轻微接触隐患（一般热缺陷）； like:

, the part has a slight contact hazard (general thermal defect);

，则该部件有重大缺陷； like:

, the part has major defects;

，则该部件为紧急缺陷； like:

, the component is an emergency defect;

2-5. According to the above steps 2-1 to 2-4, perform thermal defect diagnosis on other identified components;

Compared with the prior art, the present invention has the following obvious outstanding substantive features and significant advantages: the present invention adopts a clustering algorithm based on line object perception to identify line components, which can effectively eliminate the heat generated by cars driving on the ground in natural and complex scenes The influence of heat reflected by objects and towers on the diagnosis of line thermal defects. The invention adopts the relative temperature difference method to diagnose thermal defects of components, which can eliminate the influence of additional temperature rise caused by solar radiation, and reduce errors caused by inaccurate parameters such as detection distance, ambient temperature, humidity, and wind speed. The invention effectively improves the efficiency of detecting thermal defects of transmission lines, and can be effectively applied to vehicle-mounted or helicopter transmission line inspection services.

Description of drawings

Fig. 1 is a working environment diagram of the intelligent detection thermal defect hardware of the present invention;

Fig. 2 is the method for identifying high-voltage circuit components from the infrared grayscale image of the present invention;

Fig. 3 is a method for diagnosing thermal defects by the absolute temperature difference method from the components identified in Fig. 2;

Fig. 4 is the infrared image of the thermal defect on the wire of the high-voltage line actually collected;

Fig. 5 extracts and identifies the wire parallel wire group from the infrared image of Fig. 4;

Fig. 6 is the method for finding the highest temperature from segmental search on each wire identified in Fig. 5;

Fig. 7 is an example diagram of extending wires to diagnose thermal defects in the contact area from the actually collected infrared images.

Detailed ways

Preferred embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings. This embodiment is carried out on the premise of the technical solution of the present invention, and detailed implementation is given, but the protection scope of the present invention is not limited to the following embodiments.

As shown in Fig. 1, the applicable hardware environment of the method for intelligent detection of thermal defects of high-voltage lines by infrared images of the present invention can detect thermal defects in the working environment of helicopter inspection or vehicle inspection. The resolution of the infrared thermal imager is 320╳240 or 640╳480. The output interface of the infrared thermal imager is analog PAL/NTSC video data or digital network video stream. Each infrared grayscale image collected has a corresponding temperature image. Each time an infrared image is collected, infrared thermal defect diagnosis is realized at the same time.

As shown in Figure 2, the method for intelligently diagnosing thermal defects of high-voltage transmission lines based on infrared images has the following steps:

1. Use the line object perception clustering algorithm to identify line components. The specific steps are as follows:

1-1. The preprocessing of the collected infrared images includes: using 3╳3 median filter for noise reduction, histogram equalization processing to obtain infrared grayscale images with clear edges, and the infrared grayscale images are processed by Prewiit operators in four directions. The edge is extracted from the high-degree image, and the infrared binary image is generated by the maximum two-difference method between classes. Denoted as IB (i, j) to represent the infrared binary image; Fig. 4 is the infrared image actually collected, and Fig. 5 is the infrared binary image generated through processing;

的直线是由水平小线段组成；倾角范围在

的直线是由倾斜小线段组成；倾角范围在

的直线是由垂直小线段组成。在二值图像上按8搜索并标记水平、倾斜 、垂直小线段，定义三个链表结构Hsegment、Vsegment、Ssegment用来存放水平、垂直、倾斜小线段； 1-2. Extract horizontal, oblique and vertical small line segments on the binary image. Inclination range in

The straight line is composed of small horizontal line segments; the inclination range is in

The straight line is composed of small inclined line segments; the range of inclination is in

The straight line is composed of small vertical line segments. Press 8 to search and mark horizontal, inclined, and vertical small line segments on the binary image, and define three linked list structures Hsegment, Vsegment, and Ssegment to store horizontal, vertical, and inclined small line segments;

1-3. Calculate the slope and intercept of each horizontal, inclined, and vertical small line segment, connect the small line segments with the same slope and intercept into long straight line segments; fit the curve segment by analyzing the common endpoint of the small line segments ; Use the following structure to represent a line segment:

struct Segment{

       Point startP; //The coordinates of the starting point of the line segment

        Point endP;                                                      

      Double slope; //Line segment slope

       Double intercept; //Line segment intercept

           Int length;                                                                                              

                                                                   

}

1-4. Divide the area vertically on the binary image, and analyze the distribution density of three types of small line segments in each area; if there is an area with relatively high distribution of horizontal, vertical, and inclined line segments at the same time, it is confirmed as the tower area. In the tower image Whether there is a curve in the image to distinguish whether it is the second type of image or the third type of image, if there is no curve in the tower image, it is confirmed as the second type of infrared image; if there is a curve in the tower image, it is confirmed as the third type of infrared image;

1-5. Analyze the parallel relationship of horizontal, vertical, and inclined line segments, and cluster the line segments with approximately the same slope and different intercepts into parallel line groups; and confirm them as wire groups. Use the following structure to represent groups of parallel lines:

struct Parallel{

Int Group;

Int size;

          Point startP;

Point endP;

Double slope;

Int left_edge;

Int right_edge;

Int bottom_edge;

Int up_edge;

            }

1-6. Detect the width of the wire according to the identified position of the wire. According to the position of the insulator in the second type of infrared image (at the vertical position at the breakpoint of the two parallel broken lines) and the position in the third type of infrared image (in the direction of parallel extension at the end of the parallel wire), the approximate position of the insulator is deduced.

1-7. On the basis of generating the binary template images of the conductors, drain wires, and insulators, manage each conductor, drain wire, and insulator with a binary connected area. In order to effectively detect the contact area between the wire and other parts of the circuit, the binary connected area is extended so that it can effectively expand to the contact area between the parts. The identified components are managed using the following structure:

struct part_object {

PART_NAME part_ID;

unsigned int elements_number;

`` vector<struct Element> elements;

           //The coordinates of the rectangular area bounded by the component

int leftx, lefty;

int topx, topy;

int rightx, righty;

int bottomx, bottom;

}

2. As shown in Figure 3, on the basis of infrared image recognition of high-voltage line components, infrared thermal defect diagnosis is performed, and the specific steps are as follows:

2-1. The coordinates of each identified component are managed by BLOB in the connected area of the object. According to its pixel coordinates, the actual temperature value in the area is recorded from the infrared temperature image, and the highest temperature value is automatically found out for the wire components in blocks. , for the insulator parts, according to the connectivity, automatically find out the maximum temperature value, which is recorded as T _{max_hot} , as shown in Figure 6, the maximum temperature automatically found by using the wire block method is marked with a green color box;

2-2. Use the highest temperature pixel as the seed point, use the region growing method, and use 5 degrees lower than the highest temperature as the boundary judgment condition to generate the connected region with the highest temperature. For example, the size of the highest temperature region is approximately the same as the temperature region of the component , then the temperature of the component is normal, and the diagnosis of the component is ended; if the size of the highest temperature region is obviously smaller than the size of the temperature region of the component, the pixels outside the highest temperature region of the component are regarded as a region;

2-3. Generate two temperature connected regions on the infrared temperature image IW(i,j) for this component region: one is the connected region with the highest temperature, and the other is the connected temperature region composed of pixels other than the highest temperature. Calculate the average value AVG_T _max of the temperature in the connected region with the highest temperature, and calculate the average value AVG_T _ref of the temperature in the connected region composed of pixels other than the highest temperature, as follows:

诊断该部件热缺陷： 2-4. Basis

Diagnose the component thermal defect:

 ，则该部件有轻微接触隐患（一般热缺陷）； like:

, the part has a slight contact hazard (general thermal defect);

，则该部件有重大缺陷； like:

, the part has major defects;

，则该部件为紧急缺陷； like:

, the component is an emergency defect;

= 150%，该缺陷属于紧急热缺陷； As shown in FIG. 7 , the actually collected defective infrared image has T _max = 48 ^. C, T _ref = 18 ^. C,

= 150%, the defect is an emergency thermal defect;

2-5. According to the above steps 4-1 to 4-4, perform thermal defect diagnosis on other identified components, and use the following structure to manage the diagnosed thermal defect components:

struct tagIRFault

{

FAULT_PART_NAME IrFaultPart;

//The name of the defective part of infrared diagnosis is: NULL_FAULT: no defect; otherwise, it is the defective area as follows

                        IrFaultRect;

                     Highesttemperature;

int int DiffTempe; //relative temperature difference;

}.

Claims (3)

1. method based on infrared image intelligent diagnostics ultra-high-tension power transmission line thermal defect is characterized in that operation steps is following:

Step 1, employing are based on line object perception clustering algorithm identification circuit parts;

Step 2, the circuit parts to discerning adopt its thermal defect of relative temperature difference method diagnosis.

2. the method based on infrared image intelligent diagnostics ultra-high-tension power transmission line thermal defect as claimed in claim 1, it is following based on the concrete steps of line object perception clustering algorithm to it is characterized in that described step 1 adopts:

The boundary operator of step 1-1, band direction is handled infrared image, extracts horizontal line section, vertical line segment, oblique line section, segment of curve, generates infrared bianry image through two difference methods between maximum kind, and (i j) representes infrared bianry image to be designated as IB;

Step 1-2, on the infrared bianry image of step 1-1, calculate the slope and the intercept of each level, inclination, vertical little line segment, the little line segment that slope is identical with intercept merges and to be connected to long straight lines; Mode through analyzing the common end points of little line segment is come the matched curve section;

Step 1-3, on the infrared bianry image of step 1-2 perpendicular bisected zone, analyze three types little line segment distribution densities in each zone; As have level, vertical, the higher zone of inclined line segment distribution density simultaneously, confirm as the infrared image that comprises shaft tower; Do not comprise the straight line pole infrared image as not having curve in the shaft tower image, confirming as; There is curve in the shaft tower image, confirms as and comprise strain insulator shaft tower infrared image;

Step 1-4, with in horizontal line section and the inclined line segment, slope is approximate identical, the line segment that intercept is different classifies as sets of parallel; And confirm as the lead group;

Step 1-5, according to step 1-3 and step 1-4, comprise in the infrared image of straight line pole confirming, as to have two groups of parallel wires; Judge adjacent end points degree of closeness; Simultaneously relatively the expansion lead intersects these two groups of leads, has composite insulator in the vertical direction of intersection point; Infer the composite insulator position, as the composite insulation subassembly of identification; As have one group of parallel wire, and with the adjacent end points in shaft tower zone, press its end points of 16*8 size pixel area extension, make it effectively expand to the contact area between the circuit parts;

Step 1-6, according to step 1-3 and step 1-4; Comprise in the infrared image of strain insulator shaft tower in affirmation; As having the parallel wire group, there are the glass insulation subassembly in judgement and curve adjacent wires end points on the parallel extending direction of this end points; Infer the glass insulator position, as the glass insulation subassembly of identification;

Step 1-7, the lead with identification, drainage thread, insulator position coordinates are with independently connected region structure management.

3. the method based on intelligent diagnostics ultra-high-tension power transmission line thermal defect in the infrared image as claimed in claim 2 is characterized in that the circuit parts of said step 2 pair identification, adopts the concrete steps of its thermal defect of relative temperature difference method diagnosis following:

Step 2-1 reads each component area of identification, according to its area pixel coordinate; From infrared temperature image I W (i, j) on, read in this zone actual temperature value; Adopt piecemeal to find out maximum temperature value automatically to wire guiding member; Insulator part according to connectedness, is found out maximum temperature value automatically, be designated as T _{Max_hot}

Step 2-2, be seed points with the maximum temperature pixel; Adopt region growing method; To be lower than maximum temperature 5 degree as the border judgment condition, generate the connected region of maximum temperature, approximate consistent like the big peanut of this highest temperature region with the big peanut of this component area; Then this part temperatures is normal, then finishes diagnosis; Be significantly less than the big peanut of this component area like the big peanut of highest temperature region, as a zone, change this parts highest temperature region exterior pixel over to step 2-3;

Step 2-3, to this component area at infrared temperature image I W (i; J) go up two temperature connected regions of generation: one is the connected region of maximum temperature; Another calculates the mean value AVG_T of temperature in the maximum temperature connected region by removing the temperature connected region that the maximum temperature exterior pixel is formed _Max, calculate mean value AVG_T except that temperature in the temperature connected region of maximum temperature exterior pixel composition _Ref, do following calculating:

；

Step 2-4, foundation

are diagnosed this parts thermal defect:

As:

; Then these parts have slight contact hidden danger, promptly general thermal defect;

As:

, then these parts have significant deficiency;

As:

, then these parts are urgent defective;

Step 2-5, according to above-mentioned steps 2-1 to step 2-4, other identification components are carried out the thermal defect diagnosis.

Images