Semi-autonomous naming method for overhead transmission line photo based on YOLOv5 algorithm

Abstract

The invention relates to an overhead transmission line photo semi-autonomous naming method based on a YOLOv5 algorithm, and belongs to the field of overhead transmission line operation and maintenance. The method comprises the following steps: S1, establishing a power transmission line photo data set, and training by adopting a YOLOv5 algorithm to obtain a target detection model; s2, establishing a tower information database, reading latitude and longitude information in the picture, comparing the latitude and longitude information with latitude and longitude information of each tower in the tower information database, and calculating to obtain the tower to which the picture belongs; s3, using the trained target detection model to identify the electric power facility in the photo; and S4, if the electric power facility in the photo is an insulator string or an accessory part thereof, judging a large size side and a small size side according to latitude and longitude information, and generating a photo name by integrating a line name, a tower number and an electric power facility name.

Description

Technical Field

[0001] The present invention relates to a semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm, belonging to the field of overhead transmission line operation and maintenance. Background Art

[0002] As a daily task in the field of power transmission operation and maintenance, fine inspection of overhead transmission lines by drones will generate a large number of photos. Staff need to complete the naming of the photos according to the photo information. The photo name usually consists of three parts: line name, tower number, and power facility name. Quickly and accurately completing the naming of inspection photos is of great significance for maintaining the transmission line ledger and ensuring the safe and stable operation of the line.

[0003] The existing photo naming method manually inputs the line name and tower number of the drone inspection, and carefully plans the inspection route of the drone, so that the drone takes pictures of the power facilities on the tower in a fixed order, and then autonomously completes the photo naming work according to the photo shooting order. This method reduces the manual workload to a certain extent and improves the photo naming efficiency, but there are also some deficiencies. For example, when the user inputs the wrong line name or tower number, or the shooting order is disrupted, it will cause a large-scale error in photo naming. Summary of the Invention

[0004] The purpose of the present invention is to overcome the above deficiencies existing in the prior art, and provide a reasonably designed semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm, which realizes the semi-autonomous naming of transmission line inspection photos by using the latitude and longitude information in the photos and the YOLOv5 image recognition algorithm.

[0005] The technical solution adopted by the present invention to solve the above problems is: the semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm is characterized by including the following steps:

[0006] S1, establish a dataset of overhead transmission line photos, and train a target detection model using the YOLOv5 algorithm;

[0007] S2, establish a tower information database, read the latitude and longitude information in the photo, compare it with the latitude and longitude information of each tower in the tower information database, and calculate the tower to which the photo belongs;

[0008] S3, use the trained target detection model to identify the power facilities in the photo;

[0009] S4, if the power facility in the photo is an insulator string or its accessory parts, then determine the large and small side according to the latitude and longitude information, and generate the photo name by integrating the line name, tower number, and power facility name.

[0010] Further, S1 specifically includes the following steps:

[0011] S1-1, establish a dataset of transmission line photos;

[0012] S1-2, use the YOLOv5 algorithm to train the dataset to obtain an object detection model.

[0013] Further, the dataset of transmission line photos includes a total of eleven types of photos, namely: pole number plate, pole tower foundation, overall view of the pole tower, tower head part, earth wire fitting, insulator string, cross arm end of the insulator string, conductor end of the insulator string, hanging string, cross arm end of the hanging string, conductor end of the hanging string.

[0014] Further, S2 specifically includes the following steps:

[0015] S2-1, establish a pole tower information database;

[0016] S2-2, read the longitude and latitude data L A (Lon A , Lat A ) in the current photo;

[0017] S2-3, set the initial value of the minimum distance min_d to 100, and traverse the longitude and latitude data L i (Lon i , Lat i ) of each pole tower in the pole tower information database, where i = 1, 2,...... N, and N is the total number of pole towers in the pole tower information database, and calculate the distance d A between the point L i and the point L i , and the calculation formula is as follows:

[0018]

[0019] In the formula,

[0020] If d i < min_d, then let min_d = d i

[0021] S2-4, after traversing the pole tower information database, obtain the minimum distance value min_d. If min_d = 100, then the current data is invalid and skip the current photo; if min_d < 100, then the pole tower corresponding to the min_d value is the pole tower to which the photo belongs.

[0022] Further, the pole tower information database includes five types of data, namely: line name, pole tower number, longitude, latitude, and pole tower number on the large side.

[0023] Further, S3 specifically includes the following steps:

[0024] S3-1. Use the target detection model to identify the power facilities {E1, E2,..., E n} in the photo, where n is the total number of identified power facilities;

[0025] S3-2. If the photo only includes two categories, namely the pole number plate and the pole tower foundation, set the power facilities captured in the current photo as the pole number plate; if the photo includes two categories, namely the overall view of the pole tower and the tower head part, set the power facilities captured in the current photo as the overall view of the pole tower;

[0026] S3-3. For other cases, calculate the pixel areas {S1, S2,......, S n} occupied by each power facility {E1, E2,..., E n} in the photo;

[0027] S3-4. Compare to obtain the maximum pixel area S max , and the power facility corresponding to it is the power facility captured in the current photo.

[0028] Further, S4 specifically includes the following steps:

[0029] S4-1. Determine whether the current photo is an insulator string or its accessory parts. If so, execute S4-2 to S4-5 to determine the large and small side of the insulator string; otherwise, execute S4-6 to S4-7 to generate the photo name;

[0030] S4-2. Read the longitude and latitude L C (Lon C , Lat C ) of the current pole tower and the longitude and latitude L D (Lon D , Lat D ) of the large side pole tower from the pole tower information database;

[0031] S4-3. According to the Miller projection method, convert the longitude and latitude L A (Lon A , Lat A ) of the current photo, the longitude and latitude L C (Lon C , Lon C ) of the current pole tower, and the longitude and latitude L D (Lon D , Lat D ) of the large side pole tower into plane coordinates P A (x A , y A ), P c (xc , y c ), P D (x D , y D ), taking the longitude and latitude L A as an example, the conversion formula is as follows:

[0032]

[0033] In the formula, w = 6378245 * π * 2, h = 6378245 * π

[0034] S4-4, translate the coordinate system so that point P c is the origin. The coordinates of point P A after the coordinate system translation are P A′ (x A′ , y A′ ) and the coordinates of point P D after the coordinate system translation are P D′ (x D′ , y D′ ). The calculation formula is as follows:

[0035] x A′ = x A - xc y A′ = y A - y c

[0036] x D′ = x D - x c y D′ = y D - y c

[0037] S4-5, determine the large and small sides to which the current photo belongs based on point P A′ and point P D′ . The discrimination rule is as follows:

[0038]

[0039] If then the current photo belongs to the large side. If then the current photo belongs to the small side, where is the critical angle;

[0040] S4-6, if the power facilities captured in the photo are one of the following six: insulator string, insulator string cross-arm end, insulator string conductor end, suspension string, suspension string cross-arm end, suspension string conductor end, the user needs to determine the phase to which the current power facility belongs;

[0041] S4-7, generate the name of the current photo.

[0042] Further, if the power facilities captured in the photo are one of the following three: insulator string, cross-arm end of insulator string, conductor end of insulator string, then it is considered that the power facilities captured in the photo are an insulator string or its accessory components.

[0043] Further, the critical value

[0044] Further, the naming specification of the photo name is: line name + tower number + phase (if any) + large and small side (if any) + power facility name.

[0045] Compared with the prior art, the present invention has the following advantages: By using the YOLOv5 algorithm to train the power transmission line photo dataset to obtain a target detection model, the tower captured in the photo is located using the longitude and latitude information in the photo to be named, and the power facilities captured in the photo are obtained using the target detection model, so as to achieve the purpose of semi-autonomous naming of power transmission line inspection photos.

[0046] This application has the advantages of high automation and high accuracy rate, improves the deficiency of error-proneness in the existing photo autonomous naming method to a certain extent, effectively improves the naming efficiency of power transmission line inspection photos, and provides strong technical support for the maintenance of power transmission line ledger information. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flowchart of the semi-autonomous naming method for overhead power transmission line photos in an embodiment of the present invention.

[0048] Figure 2 is a flowchart of the algorithm for discriminating the tower to which the photo belongs in an embodiment of the present invention.

[0049] Figure 3 is a schematic diagram of the situation where a photo in an embodiment of the present invention includes both a tower foundation and a pole number plate.

[0050] Figure 4 is a schematic diagram of the situation where a photo in an embodiment of the present invention includes both the overall view of the tower and the tower head part.

[0051] Figure 5 is a schematic diagram of the pixel area occupied by power facilities in a photo in an embodiment of the present invention.

[0052] Figure 6 is a schematic diagram of the discrimination principle of the large and small sides of an insulator string and its accessory components in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The present invention will be further described in detail below in conjunction with the accompanying drawings and through embodiments. The following embodiments are explanations of the present invention, and the present invention is not limited to the following embodiments.

[0054] Embodiment

[0055] Refer to Figures 1 to 6 As shown, it should be noted that in this specification, if there are terms such as "upper", "lower", "left", "right", "middle", and "one" cited, they are only for the convenience of clear narration and are not used to limit the scope of implementation of the present invention. The change or adjustment of their relative relationship, without substantial change in the technical content, should also be regarded as the scope of implementation of the present invention.

[0056] The semi-automatic naming method for overhead transmission line photos based on the YOLOv5 algorithm in this embodiment includes the following steps:

[0057] S1. Establish a dataset of transmission line photos and train a target detection model using the YOLOv5 algorithm;

[0058] S1 specifically includes the following steps:

[0059] S1-1. Establish a dataset of transmission line photos;

[0060] S1-2. Use the YOLOv5 algorithm to train the dataset to obtain a target detection model;

[0061] The dataset of transmission line photos includes a total of eleven types of photos, namely: pole number plate, tower foundation, overall view of the tower, tower head part, earth wire fitting, insulator string, cross arm end of the insulator string, conductor end of the insulator string, hanging string, cross arm end of the hanging string, conductor end of the hanging string.

[0062] S2. Establish a tower information database, read the longitude and latitude information in the photo, compare it with the longitude and latitude information of each tower in the tower information database, and calculate the tower to which the photo belongs;

[0063] S2 specifically includes the following steps:

[0064] S2-1. Establish a tower information database;

[0065] S2-2. Read the longitude and latitude data L in the current photo A (Lon A , Lat A );

[0066] S2-3. Set the initial value of the minimum distance min_d to 100, and traverse the longitude and latitude data L of each tower in the tower information database i (Lon i , Lat i), where \(i = 1, 2,\cdots,N\), \(N\) is the total number of poles in the pole information database, and the calculation point \(L\) A and the point \(L\) i the distance \(d\) between them i is calculated as follows:

[0067]

[0068] In the formula,

[0069] If \(d\) i \(< \min_d\), then let \(\min_d = d\) i

[0070] In S2 - 4, after traversing the pole information database, the minimum distance value \(\min_d\) is obtained. If \(\min_d = 100\), the current data is invalid and the current photo is skipped; if \(\min_d < 100\), the pole corresponding to the \(\min_d\) value is the pole to which the photo belongs;

[0071] The pole information database includes five types of data, namely: line name, pole number, longitude, latitude, and pole number on the large - side of the pole.

[0072] S3. Use the trained object - detection model to identify the power facilities in the photo;

[0073] S3 specifically includes the following steps:

[0074] S3 - 1. Use the object - detection model to identify the power facilities \(\{E1, E2,\cdots,E\) n \}\) in the photo, where \(n\) is the total number of identified power facilities;

[0075] S3 - 2. If the photo only includes two categories, namely the pole number plate and the pole foundation, set the power facilities photographed in the current photo as the pole number plate; if the photo includes two categories, namely the overall view of the pole and the tower head part, set the power facilities photographed in the current photo as the overall view of the pole;

[0076] S3 - 3. For other cases, calculate the pixel area \(\{S1, S2,\cdots,S\) n \}\) occupied by each power facility \(\{E1, E2,\cdots,E\) n \}\) in the photo;

[0077] S3 - 4. Compare to obtain the maximum pixel area \(S\) max , and the power facility corresponding to it is the power facility photographed in the current photo.

[0078] S4. If the power facilities in the photo are insulator strings or their accessory components, determine the large and small side according to the longitude and latitude information, and generate the photo name by integrating the line name, pole number, and power facility name.

[0079] S4 specifically includes the following steps:

[0080] S4-1. Determine whether the current photo is an insulator string or its accessory component. If so, execute S4-2 to S4-5 to determine the large and small sides to which the insulator string belongs; otherwise, execute S4-6 to S4-7 to generate the photo name.

[0081] S4-2. Read the longitude and latitude L of the current pole from the pole information database C (Lon C , Lat C ) and the longitude and latitude L of the large-side pole D (Lon D , Lat D );

[0082] S4-3. According to the Miller projection method, convert the longitude and latitude L of the current photo A (Lon A , Lat A ), the longitude and latitude L of the current pole C (Lon C , Lat C ) and the longitude and latitude L of the large-side pole D (Lon D , Lat D ) into the plane coordinates P A (x A , y A ), P c (x c , y c ), P D (x D , y D ). Taking the longitude and latitude L A as an example, the conversion formula is as follows:

[0083]

[0084] In the formula, w = 6378245 * π * 2, h = 6378245 * π

[0085] S4-4. Translate the coordinate system so that the point P c is the origin, and find the coordinates of the point P A after the coordinate system translation, the point coordinates P A′ (x A′ , y A′ ) and the point coordinates P D after the coordinate system translation D′(x D′ , y D′ ), the calculation formula is as follows:

[0086] x A′ = x A - x c y A′ = y A - y c

[0087] x D′ = x D - x c y D′ = y D - y c

[0088] S4-5. From point P A′ and point P D′ to determine the large and small sides to which the current photo belongs. The determination rules are as follows:

[0089]

[0090] If then the current photo belongs to the large side. If then the current photo belongs to the small side. Among them is the critical angle, and the critical value

[0091] S4-6. If the power facilities photographed in the photo are one of the following six: insulator string, insulator string cross-arm end, insulator string conductor end, hanging string, hanging string cross-arm end, hanging string conductor end, the user needs to determine the phase to which the current power facility belongs;

[0092] S4-7. Generate the name of the current photo;

[0093] If the power facilities photographed in the photo are one of the following three: insulator string, insulator string cross-arm end, insulator string conductor end, it is considered that the power facilities photographed in the photo are insulator strings or their accessories.

[0094] The naming specification of the photo name is: line name + tower number + phase (if any) + large and small sides (if any) + power facility name.

[0095] Specifically, the semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm has a flowchart as Figure 1 shown, including the following steps:

[0096] The first step: Establish a dataset of transmission line photos and train a target detection model using the YOLOv5 algorithm.

[0097] Combined with the refined inspection requirements of UAVs for transmission lines, the established photo dataset of transmission lines can be divided into 11 categories: (pole number plate, pole tower foundation, overall view of pole tower, tower head part, ground wire fittings, insulator string, cross-arm end of insulator string, conductor end of insulator string, hanging string, cross-arm end of hanging string, conductor end of hanging string). For each category, more than 2,000 photos are selected, and the power facilities in the photos are manually labeled to form a training set. Then, the YOLOv5 algorithm can be used to train the object detection model.

[0098] Step 2: Establish a pole tower information database, read the longitude and latitude information in the photo, compare it with the longitude and latitude information of each pole tower in the pole tower information database, and calculate the pole tower to which the photo belongs.

[0099] Establish a pole tower information database, which can exist in any form such as an Excel file or a database file. However, the pole tower information database needs to present the following five types of data (line name, pole tower number, longitude, latitude, pole tower number on the large side). Table 1 shows a form of the pole tower information database.

[0100] Table 1 Sample of pole tower information database

[0101] Line name Pole tower number Precise latitude Precise longitude Pole tower number on the large side Hangzhou XXXX Line 050# 30.630244 119.753069 051# Hangzhou XXXX Line 051# 30.827067 119.76765 052# Hangzhou XXXX Line 052# 30.324284 119.964154 053# ... ... ... ... ...

[0102] For the photo to be named, first parse the EXIF information of the photo and read the longitude and latitude L of the current photo A (Lon A , Lat A )(The longitude and latitude information in the photo is the longitude and latitude where the UAV is located when taking this photo). To determine the line and pole tower where the current photo is taken, it is necessary to traverse the information of each pole tower in the pole tower information database one by one, and calculate the longitude and latitude L of each pole tower in the pole tower information database i (Lon i , Lat i ) and the straight-line distance between the photo longitude and latitude L A (Lon A , Lat A ), where i = 1, 2,...... N, and N is the total number of pole towers in the pole tower information database. The calculation formula is as follows:

[0103]

[0104] In the formula,

[0105] If d i < min_d, then let min_d = d i, where min_d is the minimum distance. To eliminate invalid photos, the initial value of min_d is set to 100 meters. After traversing the pole and tower information database, the minimum distance value min_d is obtained. If min_d = 100, the current data is invalid and the current photo is skipped; if min_d < 100, the line and pole and tower corresponding to this min_d value are the line and pole and tower to which the photo belongs. Figure 2 Better shows the process of determining the pole and tower to which the photo belongs.

[0106] Step 3: Use the trained object detection model to identify the power facilities in the photo.

[0107] After determining the line and pole and tower to which the photo belongs, the object detection model trained in the first step can be used to identify the power facilities {E1, E2,..., E n} in the photo, where n is the total number of identified power facilities. To deal with some special situations, the present invention formulates the following rules:

[0108] If the photo only includes two categories, namely the pole number plate and the pole and tower foundation, the target power facility photographed in the current photo is set as the pole number plate, Figure 3 shows the situation where the photo only includes two categories, namely the pole number plate and the pole and tower foundation; if the photo includes two categories, namely the overall view of the pole and tower and the tower head part, the target power facility photographed in the current photo is set as the overall view of the pole and tower, Figure 4 shows the situation where the photo includes two categories, namely the overall view of the pole and tower and the tower head part.

[0109] In addition, the present application takes the power facility with the largest pixel area in the photo as the target power facility. For this purpose, the pixel area {S1, S2,......, S n} occupied by each power facility {E1, E2,..., E n} in the photo should be calculated. Figure 5 shows the pixel area occupied by the power facilities in the photo. By comparison, the largest pixel area S max can be obtained, and the power facility corresponding to it is the target power facility photographed in the current photo.

[0110] Step 4: If the power facility in the photo is an insulator string or its accessory parts, determine the large and small side according to the longitude and latitude information, and generate the photo name by integrating the line name, pole and tower number, and power facility name.

[0111] When the power facility photographed in the photo is an insulator string or its accessory parts, it is necessary to further determine its large and small side. The determination method is to use the longitude and latitude L C (Lon C , Lat C ) of the current pole and tower, the longitude and latitude L D(Lon D , Lat D ) and the latitude and longitude L of the current photo A (Lon A , Lat A ), the angle relationship between them can be read from the pole tower information database. The latitude and longitude L of the current pole tower C (Lon C , Lat C ) and the latitude and longitude L of the pole tower on the larger side D (Lon D , Lat D ). According to the Miller projection method, the latitude and longitude L of the current photo A (Lon A , Lat A ), the latitude and longitude L of the current pole tower C (Lon C , Lat C ) and the latitude and longitude L of the pole tower on the larger side D (Lon D , Lat D ) are converted into plane coordinates P A (x A , y A ), P c (x c , y c ), P D (x D , y D ). Taking the latitude and longitude L A as an example, the conversion formula is as follows:

[0112]

[0113] In the formula, w = 6378245 * π * 2, h = 6378245 * π.

[0114] For the convenience of calculation, in this application, the coordinate system is translated, and the P c point is converted into the coordinate origin, and the coordinates of the P A point after the coordinate system translation are obtained, P A‘ (x A′ , y A′ ) and the coordinates of the P D point after the coordinate system translation, P D′ (x D′ , y D′ ). The calculation formula is as follows:

[0115] x A′ = x A - x c y A′ = y A - y c

[0116] x D′ = x D -x c y D′ = y D -y c

[0117] From point P A′ and point P D′ Discriminate the large and small sides to which the current photo belongs. The discrimination rules are as follows:

[0118]

[0119] If then the current photo belongs to the large side. If then the current photo belongs to the small side, where is the critical angle, taking 90°. Figure 6 Shows the discrimination principle of the large and small sides of the insulator string and its accessories.

[0120] If the power facilities photographed in the photo are one of the following six (insulator string, cross-arm end of insulator string, conductor end of insulator string, suspension string, cross-arm end of suspension string, conductor end of suspension string), the user also needs to discriminate the phase to which the current power facility belongs;

[0121] Finally, the user can generate the name of the transmission line photo in the following format: line name + tower number + phase (if any) + large and small sides (if any) + power facility name, such as insulator string conductor end on the small side of phase B of tower 005 of line Haida 23C9.

[0122] In addition, it should be noted that any equivalent changes or simple changes made according to the structure, features, and principles described in the inventive concept of this invention patent are included in the protection scope of this invention patent. Those skilled in the art of this invention can make various modifications, supplements, or use similar methods to replace the specific embodiments described, as long as they do not deviate from the structure of this invention or exceed the scope defined by this claim book, they should all fall within the protection scope of this invention.

Claims

1. A semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm, characterized in that: It includes the following steps: S1. Establish a dataset of transmission line photos and train a target detection model using the YOLOv5 algorithm; S2. Establish a tower information database, read the longitude and latitude information in the photo, compare it with the longitude and latitude information of each tower in the tower information database, and calculate the tower to which the photo belongs; S3. Use the trained target detection model to identify the power facilities in the photo; S4. If the power facilities in the photo are insulator strings or their accessories, then determine the large and small side according to the longitude and latitude information, and generate a photo name by integrating the line name, tower number, and power facility name.

2. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 1, characterized in that: The specific steps of S1 are as follows: S1-1. Establish a dataset of transmission line photos; S1-2. Use the YOLOv5 algorithm to train the dataset to obtain a target detection model.

3. The semi-automatic naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 2, characterized in that: The dataset of transmission line photos includes a total of eleven types of photos, namely: pole number plate, tower foundation, overall view of the tower, tower head part, ground wire fittings, insulator string, cross-arm end of insulator string, conductor end of insulator string, hanging string, cross-arm end of hanging string, conductor end of hanging string.

4. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 1, characterized in that: The specific steps of S2 are as follows: S2-1. Establish a tower information database; S2-2, Read the longitude and latitude data L in the current photo A (Lon A ,Lat A ); S2-3, set the initial value of the minimum distance min_d to 100, and traverse the longitude and latitude data L of each tower in the tower information database i (Lon i , Lat i ), where i = 1, 2,......N, and N is the total number of towers in the tower information database. Calculate the distance d A between point L i and point L i , and the calculation formula is as follows: In the formula, If d i < min_d, then let min_d = d i S2-4. Traverse the tower information database to obtain the minimum distance value min_d. If min_d = 100, the current data is invalid and the current photo is skipped; if min_d < 100, the tower corresponding to the min_d value is the tower to which the photo belongs.

5. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 4, characterized in that: The tower information database includes five types of data, namely: line name, tower number, longitude, latitude, and tower number on the large side.

6. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 1, characterized in that: The specific steps of S3 are as follows: S3-1. Use the object detection model to identify the power facilities {E1, E2,..., En} in the photo, where n is the total number of identified power facilities; n ​ S3-2. If the photo only includes two categories, namely the pole number plate and the tower foundation, set the power facilities captured in the current photo as the pole number plate; if the photo includes two categories, namely the overall view of the tower and the tower head part, set the power facilities captured in the current photo as the overall view of the tower; S3-3. For other cases, calculate the pixel area {S1, S2,......, S n} occupied by each power facility {E1, E2,..., E n} in the photo; S3-4. Compare to obtain the maximum pixel area S max , and the corresponding power facility is the power facility captured in the current photo.

7. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 1, characterized in that: The specific steps of S4 are as follows: S4-1. Determine whether the current photo is an insulator string or its accessory. If so, execute S4-2 to S4-5 to determine the large and small sides to which the insulator string belongs; otherwise, execute S4-6 to S4-7 to generate a photo name; S4-2, Read the current tower longitude and latitude L from the tower information database C (Lon C ,Lat C ) and the longitude and latitude L of the tower on the large side D (Lon D ,Lat D ); S4-3, according to the Miller projection method, the longitude and latitude of the current photo L A (Lon A ,Lat A ), the current tower longitude and latitude L C (Lon C ,Lat C ) and the longitude and latitude of the large side tower L D (Lon D ,Lat D ) is converted to plane coordinates P A (x A ,y A ), P C (x C ,y C ), P D (x D ,y D ), with longitude and latitude L A For example, the conversion formula is as follows: where w = 6378245 * π * 2, h = 6378245 * π S4-4, translate the coordinate system so that point P c is the origin, and find the coordinates of P after the coordinate system is translated A point coordinates P A′ (x A′ , y A′ ) and the coordinates of P after the coordinate system is translated D point coordinates P D′ (x D′ , y D′ ), and the calculation formula is as follows: x A′ = x A -x c y A′ = y A -y c x D′ = x D -x c y D′ = y D -y c S4-5, starting from point P A′ and point P D′ Determine whether the current photo belongs to the large or small side. The determination rules are as follows: If then the current photo belongs to the large side. If then the current photo belongs to the small side, where is the critical angle; S4-6. If the power facilities captured in the photo are one of the following six types, namely: insulator string, cross-arm end of insulator string, conductor end of insulator string, hanging string, cross-arm end of hanging string, conductor end of hanging string, the user needs to determine the phase to which the current power facility belongs; S4-7. Generate the name of the current photo.

8. The semi-automatic naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 7, characterized in that: If the power facilities captured in the photo are one of the following three types, namely: insulator string, cross-arm end of insulator string, conductor end of insulator string, then it is considered that the power facilities captured in the photo are insulator strings or their accessories.

9. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 7, characterized in that: The critical value 10. The semi-autonomous naming method for overhead transmission line photos based on the YOLOv5 algorithm according to claim 1, characterized in that: The naming specification of the photo name is: line name + tower number + power facility name; or line name + tower number + phase + power facility name; or line name + tower number + phase + large and small sides + power facility name.

Images