Foreign matter type identification method and system for power channel safety monitoring

Abstract

The invention discloses a foreign matter type identification method and system for power channel safety monitoring. The method comprises the following steps: acquiring a to-be-identified image for power channel safety monitoring; the image to be recognized is preprocessed; inputting the preprocessed to-be-recognized image into a pre-trained foreign matter type recognition model for recognition, and determining a recognition result; when it is determined that the foreign matter type with the hidden danger exists in the recognition result, the risk generated by the foreign matter is evaluated, and a target evaluation value is obtained; when it is determined that the assessment value is greater than or equal to a preset risk assessment threshold value, sending out an early warning prompt; the foreign matter type in the power channel can be timely and accurately identified, risk assessment is carried out on the foreign matter type with hidden danger, an early warning prompt is given out when the risk assessment value reaches or exceeds the threshold value, related personnel can be timely notified to carry out processing, and occurrence and expansion of potential danger are avoided.

Description

Technical Field

[0001] The present invention relates to the field of power channel safety monitoring, and in particular to a foreign body type identification method and system for power channel safety monitoring. Background Art

[0002] In today's era, facing the continuous and growing demand for electricity and the increasing challenges in the maintenance of transmission line channels, it is extremely important to actively innovate and continuously optimize the safety monitoring mechanism of power channels. In terms of the current actual situation, the current power channel safety monitoring technology has certain limitations, which will not only pose a potential threat to the stable operation of the power grid, but may even cause a series of safety accidents, which will have an adverse impact on the normal order of the social economy and will also cause a greater degree of interference to the quality of people's daily lives.

[0003] Under the existing technical conditions, there are currently the following problems in the aspect of power channel safety monitoring: First, the identification of foreign objects in the power channel is not accurate enough, and it is difficult to accurately distinguish the specific type of foreign objects; second, the foreign object risk assessment mainly relies on manual methods. In this way, the assessment results will be greatly affected by human factors, resulting in inaccurate assessment results and relatively low assessment efficiency; third, it is impossible to make effective warnings in a timely manner for foreign objects that pose hidden dangers.

[0004] In view of various circumstances, there is an urgent need for a foreign object type identification method and system for power channel safety monitoring to effectively solve the above-mentioned problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems in the above-mentioned technology to a certain extent. To this end, the first aspect of the present invention aims to propose a foreign body type identification method for power channel safety monitoring, which can timely and accurately identify the type of foreign body in the power channel, conduct risk assessment on the type of foreign body with hidden dangers, issue a warning prompt when the risk assessment value reaches or exceeds the threshold, and promptly notify relevant personnel to handle it, so as to avoid the occurrence and expansion of potential dangers.

[0006] A second aspect of the present invention aims to provide a foreign object type identification system for power channel safety monitoring.

[0007] To achieve the above object, an embodiment of the present invention proposes a method for identifying foreign object types for power channel safety monitoring, comprising:

[0008] Acquire images to be identified for power channel safety monitoring;

[0009] Preprocessing the image to be recognized;

[0010] The pre-processed image to be identified is input into a pre-trained foreign body type identification model for identification, and the identification result is determined;

[0011] When it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain a target evaluation value;

[0012] When it is determined that the target assessment value is greater than or equal to the preset risk assessment threshold, an early warning prompt is issued.

[0013] Preferably, preprocessing the image to be identified includes:

[0014] Performing image denoising on the image to be identified to obtain a denoised image;

[0015] The denoised image is enhanced to obtain a preprocessed image to be recognized.

[0016] Preferably, performing image enhancement on the denoised image to obtain a preprocessed image to be identified includes:

[0017] Performing fuzzy area recognition on the denoised image to obtain a plurality of initial target fuzzy areas;

[0018] Performing clarity calculation on the plurality of initial target blur regions to obtain clarity values ​​corresponding to the plurality of initial target blur regions;

[0019] Comparing the clarity value corresponding to the initial target blurred area with a preset clarity threshold, and when it is determined that the clarity value corresponding to the initial target blurred area is less than the preset clarity threshold, taking the initial target blurred area as the final target blurred area;

[0020] Take any final target fuzzy area as the area to be enhanced;

[0021] Obtain the HSV value of each pixel in the area to be enhanced as the first HSV value;

[0022] Performing Gaussian blur processing on each pixel point in the area to be enhanced, and obtaining the HSV value of each pixel point in the area to be enhanced after the Gaussian blur processing as a second HSV value;

[0023] Calculate the difference between the first HSV value and the second HSV value corresponding to each pixel point in the area to be enhanced as the pixel adjustment value;

[0024] Based on the pixel adjustment value, a preset pixel adjustment value-enhancement coefficient data table is searched to determine the enhancement coefficient corresponding to each pixel point;

[0025] Perform image enhancement on each pixel based on the enhancement coefficient corresponding to each pixel to obtain the final target blurred area after enhancement;

[0026] Traverse all final target fuzzy areas to obtain several enhanced final fuzzy areas;

[0027] Based on several enhanced final target blurred regions, the denoised image is enhanced to obtain a preprocessed image to be identified.

[0028] Preferably, the fuzzy area recognition is performed on the denoised image to obtain a plurality of initial target fuzzy areas, including:

[0029] Take any denoised image as the target image; divide the target image evenly into several regions;

[0030] An area is randomly selected as the target area; the gradient direction of each pixel point in the target area is obtained, and the gradient direction evaluation value of each gradient direction in the target area is determined to obtain a plurality of gradient direction evaluation values;

[0031] The gradient direction corresponding to the maximum value of the gradient direction evaluation is taken as the target gradient direction corresponding to the target area;

[0032] Obtain the grayscale value of each pixel in the target gradient direction in the target area;

[0033] Determine the smoothing value corresponding to the target area based on the grayscale value of each pixel point in the target gradient direction;

[0034] Determine the blur level value of the target area based on the maximum value and the smoothing value of the gradient direction evaluation;

[0035] Comparing the blur level value with a preset blur level threshold, and taking the target area corresponding to when the blur level value is greater than or equal to the preset blur level threshold as the initial target blur area;

[0036] Traverse all areas to obtain several initial target fuzzy areas corresponding to the target image;

[0037] All denoised images are traversed to obtain several initial target blurred regions corresponding to the denoised images.

[0038] Preferably, performing clarity calculation on the plurality of initial target blurred areas to obtain clarity values ​​corresponding to the plurality of initial target blurred areas includes:

[0039] Take an initial target blurred area at random and obtain the brightness value of each pixel in the initial target blurred area;

[0040] Calculate the brightness mean of each pixel in the initial target blur area to obtain the target brightness mean;

[0041] Calculate the difference between the maximum brightness value and the target brightness mean value in the initial target blurred area as the first brightness difference corresponding to the initial target blurred area;

[0042] Calculate the difference between the minimum brightness value in the initial target blurred area and the target brightness mean as the second brightness difference corresponding to the initial target blurred area;

[0043] Calculate the square sum of the first brightness difference value and the second brightness difference value to obtain a clarity value corresponding to the initial target blurred area;

[0044] Each initial target fuzzy area is traversed to obtain the clarity values ​​corresponding to several initial target fuzzy areas.

[0045] Preferably, the training method of the foreign body type recognition model includes:

[0046] Obtain a foreign body type recognition training set;

[0047] Inputting the foreign body type recognition training set into a preset neural network model for training to obtain an initial foreign body type recognition model;

[0048] Optimizing the parameters of the initial foreign body type recognition model using a cross-validation method to obtain an optimized initial foreign body type recognition model;

[0049] Obtain a foreign body type recognition test set;

[0050] The optimized initial foreign body type recognition model is tested based on the foreign body type recognition test set. When the test result is qualified, the trained foreign body type recognition model is obtained.

[0051] Preferably, when it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain a target evaluation value, including:

[0052] A Cartesian rectangular coordinate system is established according to the image to be identified when there is a foreign object type with hidden dangers in the identification result, and the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object are determined;

[0053] Determine whether there is overlap between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0054] If there is overlap between the power channel and the foreign object, the overlapping area of ​​the power channel and the foreign object is determined based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0055] Obtaining a preset overlapping region judgment threshold corresponding to the foreign body type, and taking a ratio of the overlapping region area to the preset overlapping region judgment threshold as a first evaluation value;

[0056] If there is no overlap between the power channel and the foreign object, determining the shortest distance between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0057] Obtaining a preset distance judgment threshold corresponding to the foreign object type, and taking a ratio of the preset distance judgment threshold to the shortest distance as a second evaluation value;

[0058] The first evaluation value and the second evaluation value are used as target evaluation values.

[0059] Preferably, when it is determined that the target assessment value is greater than or equal to a preset risk assessment threshold, a warning prompt is issued, including:

[0060] If there is overlap between the power channel and the foreign object, the first evaluation value included in the target evaluation value is compared with the preset risk evaluation threshold, and when it is determined that the first evaluation value is greater than or equal to the preset risk evaluation threshold, an early warning prompt is issued;

[0061] If there is no overlap between the power channel and the foreign object, the second evaluation value included in the target evaluation value is compared with the preset risk assessment threshold, and when it is determined that the second evaluation value is greater than or equal to the preset risk assessment threshold, an early warning prompt is issued.

[0062] Preferably, the types of foreign matter include surrounding illegal buildings, plastic greenhouses and temporary construction.

[0063] A second aspect of the present invention provides a foreign object type identification system for power channel safety monitoring, comprising:

[0064] An acquisition module, used for acquiring images to be identified for power channel safety monitoring;

[0065] A preprocessing module, used for preprocessing the image to be identified;

[0066] The recognition module is used to input the pre-processed image to be recognized into the pre-trained foreign body type recognition model for recognition and determine the recognition result;

[0067] An evaluation module is used to evaluate the risk caused by the foreign matter when it is determined that there is a foreign matter type with hidden dangers in the identification result, and obtain a target evaluation value;

[0068] The early warning module is used to issue an early warning prompt when it is determined that the target assessment value is greater than or equal to a preset risk assessment threshold.

[0069] The present invention discloses a foreign body type recognition method and system for power channel safety monitoring. By timely identifying the type of foreign bodies and assessing the risks, the safe operation of the power channel can be effectively guaranteed, and the accident risks caused by foreign bodies can be reduced. The image to be identified is acquired and pre-processed, thereby ensuring the data quality of the input model and improving the accuracy of foreign body type recognition. When the risk assessment value reaches or exceeds the threshold, an early warning prompt is issued, so that relevant personnel can take countermeasures in time to avoid the expansion of potential hazards. It is helpful to realize the intelligent and efficient management of the power channel and reduce the pressure and cost of manual inspections. It can identify and warn foreign bodies before they cause greater harm, thereby buying time and initiative for subsequent processing.

[0070] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0071] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0073] Figure 1 is a flow chart of a method for identifying foreign matter types for power channel safety monitoring according to an embodiment of the present invention;

[0074] Figure 2 is a flowchart of preprocessing the image to be recognized according to an embodiment of the present invention;

[0075] Figure 3 The invention is a block diagram of a foreign object type identification system for power channel safety monitoring according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0077] Example 1

[0078] like Figure 1 As shown, a method for identifying foreign body types for power channel safety monitoring includes steps S1-S5:

[0079] S1: Acquire the image to be identified for power channel safety monitoring;

[0080] S2: preprocessing the image to be identified;

[0081] S3: input the preprocessed image to be identified into a pre-trained foreign body type identification model for identification, and determine the identification result;

[0082] S4: When it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain a target evaluation value;

[0083] S5: When it is determined that the target assessment value is greater than or equal to the preset risk assessment threshold, a warning prompt is issued.

[0084] In this embodiment, the methods for acquiring the images to be identified for power channel safety monitoring include drone aerial photography and satellite remote sensing photography.

[0085] In this embodiment, preprocessing the image to be identified includes denoising and enhancing the image.

[0086] The working principle of the above technical solution is: first, through step S1, the image to be identified for power channel safety monitoring is obtained, which is the basic data source of the whole process. Then, in step S2, the image to be identified is preprocessed, such as denoising, enhancement, etc., to improve the quality and availability of the image, so as to prepare for subsequent accurate identification. Then in step S3, the preprocessed image to be identified is input into the pre-trained foreign body type identification model, and the model's powerful recognition ability is used to determine the specific identification result. In step S4, when it is determined that there is a foreign body type with hidden dangers in the identification result, the risk generated by the foreign body will be further evaluated to obtain a target evaluation value to measure the risk level brought by the foreign body. Finally, in step S5, once it is determined that the target evaluation value is greater than or equal to the preset risk assessment threshold, it means that the risk has reached the level that requires early warning. At this time, an early warning prompt will be issued so that relevant personnel can take countermeasures in time to ensure the safe operation of the power channel. The entire technical solution realizes the effective identification of foreign body types in the power channel and the timely early warning and processing of risks through such a series of orderly steps and operations.

[0087] The beneficial effects of the above technical solution are: by timely identifying the type of foreign matter and assessing the risk, it can effectively ensure the safe operation of the power channel and reduce the risk of accidents caused by foreign matter; obtain the image to be identified and pre-process it, ensure the data quality of the input model, and improve the accuracy of foreign matter type identification; when the risk assessment value reaches or exceeds the threshold, an early warning prompt is issued, so that relevant personnel can take timely response measures to avoid the expansion of potential hazards; it helps to realize the intelligent and efficient management of the power channel and reduce the pressure and cost of manual inspections; it can identify and warn before foreign matter causes greater harm, so as to gain time and initiative for subsequent processing.

[0088] Example 2

[0089] like Figure 2 As shown, the image to be identified is preprocessed, including S21-S22:

[0090] S21: performing image denoising on the image to be identified to obtain a denoised image;

[0091] S22: Perform image enhancement on the denoised image to obtain a pre-processed image to be recognized.

[0092] The working principle of the above technical solution is: in step S21, the acquired image to be identified is processed by a specific image denoising algorithm, the purpose of which is to remove the noise interference that may exist in the image, so that the image becomes clearer, so that the features can be extracted and the type of foreign matter can be identified more accurately later. After the denoising is completed, step S22 is entered, and the denoised image is further processed by image enhancement technology to enhance the image's contrast, brightness and other features, so that the key information in the image is more prominent and the useful details of the image are strengthened, so that when the enhanced image is input into the foreign matter type identification model, the recognition accuracy and reliability of the model can be improved, and finally more accurate and effective foreign matter type identification and subsequent risk assessment operations can be achieved.

[0093] The beneficial effects of the above technical solution are: image denoising can remove noise interference in the image, making the image clearer and more accurate, which is helpful for subsequent recognition and analysis; image enhancement can highlight important features in the image, making key information such as foreign objects more obvious, and improving the accuracy of subsequent recognition; providing better data for the foreign object type recognition model, which is conducive to better learning and recognition of the model, thereby improving the performance and reliability of the entire system; clear and feature-rich images can reduce the occurrence of misjudgments caused by image quality problems; enabling the system to better adapt to images to be recognized acquired in different environments, thereby improving the robustness of the system.

[0094] Example 3

[0095] Performing image enhancement on the denoised image to obtain a preprocessed image to be recognized, including:

[0096] Performing fuzzy area recognition on the denoised image to obtain a plurality of initial target fuzzy areas;

[0097] Performing clarity calculation on the plurality of initial target blur regions to obtain clarity values ​​corresponding to the plurality of initial target blur regions;

[0098] Comparing the clarity value corresponding to the initial target blurred area with a preset clarity threshold, and when it is determined that the clarity value corresponding to the initial target blurred area is less than the preset clarity threshold, taking the initial target blurred area as the final target blurred area;

[0099] Take any final target fuzzy area as the area to be enhanced;

[0100] Obtain the HSV value of each pixel in the area to be enhanced as the first HSV value;

[0101] Performing Gaussian blur processing on each pixel point in the area to be enhanced, and obtaining the HSV value of each pixel point in the area to be enhanced after the Gaussian blur processing as a second HSV value;

[0102] Calculate the difference between the first HSV value and the second HSV value corresponding to each pixel point in the area to be enhanced as the pixel adjustment value;

[0103] Based on the pixel adjustment value, a preset pixel adjustment value-enhancement coefficient data table is searched to determine the enhancement coefficient corresponding to each pixel point;

[0104] Perform image enhancement on each pixel based on the enhancement coefficient corresponding to each pixel to obtain the final target blurred area after enhancement;

[0105] Traverse all final target fuzzy areas to obtain several enhanced final fuzzy areas;

[0106] Based on several enhanced final target blurred regions, the denoised image is enhanced to obtain a preprocessed image to be identified.

[0107] In this embodiment, the fuzzy region recognition of the denoised image is mainly performed from the aspect of gradient structure. Because the denoising process will inevitably cause a certain degree of blurring to the image, the fuzzy region recognition is performed on the denoised image first.

[0108] In this embodiment, the clarity calculation is performed on the plurality of initial target blurred areas, which is to perform a further clarity calculation on the screened initial target blurred areas for secondary screening, thereby improving the enhancement effect and avoiding excessive enhancement.

[0109] The working principle of the above technical solution is: first, identify the blurred area of ​​the denoised image, find several initial target blurred areas that may be blurred, and then calculate the clarity values ​​of these areas. Compare the clarity value with the preset clarity threshold, and determine those areas with too low clarity values ​​as the final target blurred areas. Then, select any one of these final target blurred areas as the area to be enhanced, obtain the original HSV value of the pixel point in the area, and obtain the new HSV value after Gaussian blur processing, and calculate the difference between the two to obtain the pixel adjustment value. According to the pixel adjustment value, the corresponding enhancement coefficient is determined from the preset data table, and the image is enhanced for each pixel point using the enhancement coefficient to obtain the enhanced final target blurred area. Repeat this process to traverse all the final target blurred areas, obtain multiple enhanced final blurred areas, and finally complete the enhancement processing of the entire denoised image based on these enhanced areas, so as to obtain the pre-processed image to be identified. The purpose of this is to enhance the blurred area in the image in a targeted manner and improve the image quality, so as to better perform operations such as foreign body type identification in the future.

[0110] The beneficial effects of the above technical solution are: first, the blurred area of ​​the denoised image is identified and the clarity is calculated, and the final target blurred area is determined by comparing with the preset threshold. Then, for any final target blurred area, the HSV value of the pixel point is obtained and Gaussian blur processing is performed to obtain a new HSV value, the difference between the two is calculated as the pixel adjustment value, the enhancement coefficient is determined by querying the data table, and then the image is enhanced according to the enhancement coefficient, and finally all the final target blurred areas are traversed to complete the overall enhancement of the denoised image; the blurred area can be enhanced in a targeted manner to improve the clarity and recognizability of these areas in the image; the HSV value and Gaussian blur are used to make the enhancement effect more in line with the image characteristics and visual requirements; the quality of the pre-processed image is improved as a whole, providing a better data basis for subsequent analysis and recognition, which helps to improve the accuracy and reliability of the system.

[0111] Example 4

[0112] The fuzzy region is identified on the denoised image to obtain a number of initial target fuzzy regions, including:

[0113] Take any denoised image as the target image; divide the target image evenly into several regions;

[0114] An area is randomly selected as the target area; the gradient direction of each pixel point in the target area is obtained, and the gradient direction evaluation value of each gradient direction in the target area is determined to obtain a plurality of gradient direction evaluation values;

[0115] The gradient direction corresponding to the maximum value of the gradient direction evaluation is taken as the target gradient direction corresponding to the target area;

[0116] Obtain the grayscale value of each pixel in the target gradient direction in the target area;

[0117] Determine the smoothing value corresponding to the target area based on the grayscale value of each pixel point in the target gradient direction;

[0118] Determine the blur level value of the target area based on the maximum value and the smoothing value of the gradient direction evaluation;

[0119] Comparing the blur level value with a preset blur level threshold, and taking the target area corresponding to when the blur level value is greater than or equal to the preset blur level threshold as the initial target blur area;

[0120] Traverse all areas to obtain several initial target fuzzy areas corresponding to the target image;

[0121] All denoised images are traversed to obtain several initial target blurred regions corresponding to the denoised images.

[0122] In this embodiment, the gradient direction of each pixel in the target area is obtained, and the gradient direction evaluation value of each gradient direction in the target area is determined to obtain a plurality of gradient direction evaluation values;

[0123] The gradient direction may be given to include 8 directions, namely, up, down, left, right, upper left, lower left, upper right and lower right; the gradient direction of each pixel point that is closest to the 8 given gradient directions is used as the gradient evaluation direction of the pixel point, the number of pixels in each gradient direction of the 8 gradient directions is obtained, and the number of pixels in each gradient direction is used as the gradient direction evaluation value.

[0124] In this embodiment, determining the smoothing value corresponding to the target area based on the grayscale value of each pixel point in the target gradient direction includes:

[0125]

[0126] Among them, P i represents the smoothing value corresponding to the i-th target region; n represents the total number of all adjacent pixel combinations in the target gradient direction in the i-th target region; m represents the m-th group of adjacent pixel points in the target gradient direction in the i-th target region; Q i,m Represents the absolute value of the grayscale difference between the mth group of adjacent pixels in the target gradient direction in the i-th target area; Represents the mean of the absolute values ​​of the differences between all adjacent pixel combinations in the target gradient direction in the i-th target region.

[0127] In this embodiment, based on the maximum value and the smoothing value of the gradient direction evaluation, the blur level value of the target area is determined, including:

[0128] The ratio of the maximum value of the gradient direction evaluation to the smoothing value is used as the blur degree value of the target area.

[0129] The working principle of the above technical solution is: first, a denoised image is selected as the target image, and it is evenly divided into several regions. Then, for each region, the gradient direction of the pixel points therein is obtained and the corresponding gradient direction evaluation value is calculated, and the gradient direction with the largest evaluation value is found as the target gradient direction. Then the grayscale value of the pixel points in the target gradient direction is obtained, and the smoothing value of the target area is determined based on these grayscale values. The blur degree value of the target area is then determined by combining the maximum gradient direction evaluation value and the smoothing value. The blur degree value is compared with the preset blur degree threshold, and if it is greater than or equal to the threshold, the target area is determined as the initial target blur area. By traversing all areas, all initial target blur areas of the target image can be obtained, and finally, by traversing all denoised images, the initial target blur areas corresponding to all denoised images can be obtained. In this way, the areas in the image that may be blurred can be identified more accurately, preparing for subsequent targeted enhancement processing.

[0130] The beneficial effects of the above technical solution are: first, the denoised image is divided into several areas, and then the gradient direction and evaluation value of the pixel points in each area are obtained to determine the target gradient direction, and then the grayscale value of the pixel points in this direction is obtained and the smoothing value is calculated accordingly, and the blur degree value is calculated in combination with the maximum value of the gradient direction evaluation, and compared with the preset threshold to determine whether it is the initial target blur area, and finally all areas and denoised images are traversed to obtain all initial target blur areas; through the detailed division of the image and the analysis of each area, the blur area can be accurately found, thereby improving the accuracy of recognition; the use of information such as gradient direction to evaluate the blur degree has good scientificity and objectivity; these blur areas can be subsequently processed in a targeted manner, thereby effectively improving the overall quality of the image and providing a better data basis for further applications; this method of step-by-step analysis by region can improve processing efficiency and the stability of the effect.

[0131] Example 5

[0132] The clarity calculation is performed on the plurality of initial target blurred areas to obtain clarity values ​​corresponding to the plurality of initial target blurred areas, including:

[0133] Take an initial target blurred area at random and obtain the brightness value of each pixel in the initial target blurred area;

[0134] Calculate the brightness mean of each pixel in the initial target blur area to obtain the target brightness mean;

[0135] Calculate the difference between the maximum brightness value and the target brightness mean value in the initial target blurred area as the first brightness difference corresponding to the initial target blurred area;

[0136] Calculate the difference between the minimum brightness value in the initial target blurred area and the target brightness mean as the second brightness difference corresponding to the initial target blurred area;

[0137] Calculate the square sum of the first brightness difference value and the second brightness difference value to obtain a clarity value corresponding to the initial target blurred area;

[0138] Each initial target fuzzy area is traversed to obtain the clarity values ​​corresponding to several initial target fuzzy areas.

[0139] The working principle of the above technical solution is: first, for any initial target blurred area, obtain the brightness value of each pixel therein, and obtain the target brightness mean by calculating the average of these brightness values. Then calculate the difference between the maximum brightness value and the target brightness mean (first brightness difference) and the difference between the minimum brightness value and the target brightness mean (second brightness difference). Then, the clarity of the initial target blurred area is measured by calculating the sum of the squares of these two brightness differences. The larger the sum of squares, the greater the difference in brightness distribution and the lower the possible clarity. Finally, by traversing all the initial target blurred areas, the clarity value corresponding to each initial target blurred area is obtained, so that these blurred areas can be judged and processed according to the clarity value later.

[0140] The beneficial effects of the above technical solution are: by calculating the brightness mean, the difference between the maximum brightness and the mean, the difference between the minimum brightness and the mean, etc., the brightness distribution in the initial target blurred area is considered more comprehensively, making the calculation of clarity more reasonable and accurate; the clarity value is obtained by a specific calculation method, which can effectively quantify the clarity of the blurred area and provide a basis for subsequent precise processing; this method is simple and direct, easy to implement and understand, and can also guarantee calculation efficiency to a certain extent; it helps to accurately distinguish areas with different degrees of blur, so as to take more appropriate enhancement measures to improve image quality.

[0141] Example 6

[0142] The training method of the foreign body type recognition model includes:

[0143] Obtain a foreign body type recognition training set;

[0144] Inputting the foreign body type recognition training set into a preset neural network model for training to obtain an initial foreign body type recognition model;

[0145] Optimizing the parameters of the initial foreign body type recognition model using a cross-validation method to obtain an optimized initial foreign body type recognition model;

[0146] Obtain a foreign body type recognition test set;

[0147] The optimized initial foreign body type recognition model is tested based on the foreign body type recognition test set. When the test result is qualified, the trained foreign body type recognition model is obtained.

[0148] The beneficial effects of the above technical solution are: by obtaining a special training set for training, the model can learn the characteristics and patterns of the foreign body type, laying the foundation for accurate identification; the neural network model has strong learning and generalization capabilities and can adapt to a variety of different foreign body types; the cross-validation method optimizes parameters, which helps to improve the accuracy and reliability of the model; obtaining a test set for testing can verify the actual effect of the model and ensure its effectiveness and accuracy in practical applications; the overall solution helps to develop an efficient and accurate foreign body type identification model, which can be applied to various scenarios where foreign bodies need to be identified, improving safety and work efficiency, etc.

[0149] Example 7

[0150] When it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain the target evaluation value, including:

[0151] A Cartesian rectangular coordinate system is established according to the image to be identified when there is a foreign object type with hidden dangers in the identification result, and the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object are determined;

[0152] Determine whether there is overlap between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0153] If there is overlap between the power channel and the foreign object, the area of ​​the overlap region between the power channel and the foreign object is determined based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0154] Obtaining a preset overlapping region judgment threshold corresponding to the foreign body type, and taking a ratio of the overlapping region area to the preset overlapping region judgment threshold as a first evaluation value;

[0155] If there is no overlap between the power channel and the foreign object, determining the shortest distance between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object;

[0156] Obtaining a preset distance judgment threshold corresponding to the foreign object type, and taking a ratio of the preset distance judgment threshold to the shortest distance as a second evaluation value;

[0157] The first evaluation value and the second evaluation value are used as target evaluation values.

[0158] The working principle of the above technical solution is: when it is determined that there is a type of foreign object with hidden dangers, first establish a Cartesian rectangular coordinate system based on the image to be identified, and clarify the coordinate information of the power channel and the foreign object therein. Then determine whether the power channel and the foreign object overlap. If there is overlap, calculate the area of ​​the overlapping area based on the coordinate information, and compare it with the preset overlapping area judgment threshold corresponding to the foreign object type to obtain a first evaluation value; if there is no overlap, calculate the shortest distance between the power channel and the foreign object based on the coordinates, and then compare it with the preset distance judgment threshold corresponding to the foreign object type to obtain a second evaluation value. Finally, the first evaluation value and the second evaluation value are used together as the target evaluation value. This evaluation value can reflect the degree of risk posed by the foreign object to the power channel, so that corresponding countermeasures can be taken according to the evaluation value in the future.

[0159] The beneficial effects of the above technical solution are: by establishing a coordinate system and other means, the relationship between the power channel and foreign objects can be accurately analyzed, including overlap and distance; for overlap, the first evaluation value is obtained by calculating the area of ​​the overlapping region and comparing it with the threshold, and the risk level can be accurately quantified; for non-overlap, the shortest distance is calculated and compared with the threshold to obtain the second evaluation value, which can also be used to evaluate the risk well; by comprehensively considering the overlap and distance, the target evaluation value obtained is more comprehensive and accurate, which helps to judge the risk caused by foreign objects more scientifically; this detailed evaluation method is conducive to taking targeted measures to deal with foreign object risks and ensure the safe operation of the power channel.

[0160] Example 8

[0161] When the target assessment value is determined to be greater than or equal to the preset risk assessment threshold, an early warning prompt is issued, including:

[0162] If there is overlap between the power channel and the foreign object, the first evaluation value included in the target evaluation value is compared with the preset risk evaluation threshold, and when it is determined that the first evaluation value is greater than or equal to the preset risk evaluation threshold, an early warning prompt is issued;

[0163] If there is no overlap between the power channel and the foreign object, the second evaluation value included in the target evaluation value is compared with the preset risk assessment threshold, and when it is determined that the second evaluation value is greater than or equal to the preset risk assessment threshold, an early warning prompt is issued.

[0164] The working principle of the above technical solution is: after obtaining the target evaluation value, different judgments and processing are performed according to the positional relationship between the power channel and the foreign object. If the power channel overlaps with the foreign object, the first evaluation value is specifically compared with the preset risk assessment threshold. When the first evaluation value reaches or exceeds the preset risk assessment threshold, an early warning is issued, which indicates that the risk caused by the overlap is relatively high at this time; and if the power channel does not overlap with the foreign object, the second evaluation value is compared with the preset risk assessment threshold. When the second evaluation value reaches or exceeds the preset risk assessment threshold, an early warning is issued, which means that although there is no overlap at this time, the risk caused by the close distance is also high. Through this classification and comparison method, it is possible to more accurately judge whether it is necessary to issue an early warning according to the specific situation to ensure the safe operation of the power channel.

[0165] The beneficial effects of the above technical solution are: it can timely detect and warn of high-risk situations, especially when the target assessment value reaches or exceeds the preset risk assessment threshold, ensuring that relevant personnel can be quickly informed; for different situations where the power channel overlaps or does not overlap with foreign objects, the assessment value and the threshold are compared respectively, making the early warning mechanism more precise and accurate; it effectively improves the monitoring ability of foreign object risks in the power channel, which is conducive to taking measures in advance to reduce potential hazards and ensure the safe and stable operation of the power system; this situation-based processing method can avoid false alarms or missed alarms, and improve the reliability and pertinence of the early warning; it helps relevant departments or personnel to make timely response decisions and reduce losses and adverse effects caused by foreign objects.

[0166] Example 9

[0167] The types of foreign objects mentioned include surrounding illegal buildings, plastic greenhouses and temporary construction.

[0168] The embodiment of the second aspect of the present invention provides a foreign object type identification system for power channel safety monitoring, including:

[0169] An acquisition module, used for acquiring images to be identified for power channel safety monitoring;

[0170] A preprocessing module, used for preprocessing the image to be identified;

[0171] The recognition module is used to input the pre-processed image to be recognized into the pre-trained foreign body type recognition model for recognition and determine the recognition result;

[0172] An evaluation module is used to evaluate the risk caused by the foreign matter when it is determined that there is a foreign matter type with hidden dangers in the identification result, and obtain a target evaluation value;

[0173] The early warning module is used to issue an early warning prompt when it is determined that the target assessment value is greater than or equal to a preset risk assessment threshold.

[0174] The working principle of the above technical solution is: the acquisition module acquires the image to be identified, the preprocessing module processes the image appropriately to optimize the subsequent identification, and the identification module uses the trained model to accurately identify the type of foreign matter in the processed image. Then, after the evaluation module finds the type of foreign matter with hidden dangers, it quantitatively evaluates the risks brought by the foreign matter to obtain the target evaluation value. Finally, the early warning module compares the target evaluation value with the preset risk assessment threshold, and issues an early warning prompt in time when the target evaluation value reaches or exceeds the threshold, so as to remind relevant personnel to take countermeasures, thereby realizing the monitoring and protection of the safety of the power channel.

[0175] The beneficial effects of the above technical solution are: by timely identifying the type of foreign matter and assessing the risk, it can effectively ensure the safe operation of the power channel and reduce the risk of accidents caused by foreign matter; obtain the image to be identified and pre-process it, ensure the data quality of the input model, and improve the accuracy of foreign matter type identification; when the risk assessment value reaches or exceeds the threshold, an early warning prompt is issued, so that relevant personnel can take timely response measures to avoid the expansion of potential hazards; it helps to realize the intelligent and efficient management of the power channel and reduce the pressure and cost of manual inspections; it can identify and warn before foreign matter causes greater harm, so as to gain time and initiative for subsequent processing.

[0176] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for identifying foreign body types for power channel safety monitoring, characterized in that: include: Acquire images to be identified for power channel safety monitoring; Preprocessing the image to be recognized; The pre-processed image to be identified is input into a pre-trained foreign body type identification model for identification, and the identification result is determined; When it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain a target evaluation value; When it is determined that the target assessment value is greater than or equal to the preset risk assessment threshold, an early warning prompt is issued.

2. The method for identifying foreign matter types for power channel safety monitoring according to claim 1, characterized in that: Preprocessing the image to be identified includes: Performing image denoising on the image to be identified to obtain a denoised image; The denoised image is enhanced to obtain a preprocessed image to be recognized.

3. The method for identifying foreign matter types for power channel safety monitoring according to claim 2, characterized in that: Performing image enhancement on the denoised image to obtain a preprocessed image to be recognized, including: Performing fuzzy area recognition on the denoised image to obtain a plurality of initial target fuzzy areas; Performing clarity calculation on the plurality of initial target blur regions to obtain clarity values ​​corresponding to the plurality of initial target blur regions; Comparing the clarity value corresponding to the initial target blurred area with a preset clarity threshold, and when it is determined that the clarity value corresponding to the initial target blurred area is less than the preset clarity threshold, taking the initial target blurred area as the final target blurred area; Take any final target fuzzy area as the area to be enhanced; Obtain the HSV value of each pixel in the area to be enhanced as the first HSV value; Performing Gaussian blur processing on each pixel point in the area to be enhanced, and obtaining the HSV value of each pixel point in the area to be enhanced after the Gaussian blur processing as a second HSV value; Calculate the difference between the first HSV value and the second HSV value corresponding to each pixel point in the area to be enhanced as the pixel adjustment value; Based on the pixel adjustment value, a preset pixel adjustment value-enhancement coefficient data table is searched to determine the enhancement coefficient corresponding to each pixel point; Perform image enhancement on each pixel based on the enhancement coefficient corresponding to each pixel to obtain the final target blurred area after enhancement; Traverse all final target fuzzy areas to obtain several enhanced final fuzzy areas; Based on several enhanced final target blurred regions, the denoised image is enhanced to obtain a preprocessed image to be identified.

4. The method for identifying foreign matter types for power channel safety monitoring according to claim 3, characterized in that: The fuzzy region is identified on the denoised image to obtain a number of initial target fuzzy regions, including: Take any denoised image as the target image; divide the target image evenly into several regions; An area is randomly selected as the target area; the gradient direction of each pixel point in the target area is obtained, and the gradient direction evaluation value of each gradient direction in the target area is determined to obtain a plurality of gradient direction evaluation values; The gradient direction corresponding to the maximum value of the gradient direction evaluation is taken as the target gradient direction corresponding to the target area; Obtain the grayscale value of each pixel in the target gradient direction in the target area; Determine the smoothing value corresponding to the target area based on the grayscale value of each pixel point in the target gradient direction; Determine the blur level value of the target area based on the maximum value and the smoothing value of the gradient direction evaluation; Comparing the blur level value with a preset blur level threshold, and taking the target area corresponding to when the blur level value is greater than or equal to the preset blur level threshold as the initial target blur area; Traverse all areas to obtain several initial target fuzzy areas corresponding to the target image; All denoised images are traversed to obtain several initial target blurred regions corresponding to the denoised images.

5. The method for identifying foreign matter types for power channel safety monitoring according to claim 4, characterized in that: The clarity calculation is performed on the plurality of initial target blurred areas to obtain clarity values ​​corresponding to the plurality of initial target blurred areas, including: Take an initial target blurred area at random and obtain the brightness value of each pixel in the initial target blurred area; Calculate the brightness mean of each pixel in the initial target blur area to obtain the target brightness mean; Calculate the difference between the maximum brightness value and the target brightness mean value in the initial target blurred area as the first brightness difference corresponding to the initial target blurred area; Calculate the difference between the minimum brightness value in the initial target blurred area and the target brightness mean as the second brightness difference corresponding to the initial target blurred area; Calculate the square sum of the first brightness difference value and the second brightness difference value to obtain a clarity value corresponding to the initial target blurred area; Each initial target fuzzy area is traversed to obtain the clarity values ​​corresponding to several initial target fuzzy areas.

6. The method for identifying foreign matter types for power channel safety monitoring according to claim 1, characterized in that: The training method of the foreign body type recognition model includes: Obtain a foreign body type recognition training set; Inputting the foreign body type recognition training set into a preset neural network model for training to obtain an initial foreign body type recognition model; Optimizing the parameters of the initial foreign body type recognition model using a cross-validation method to obtain an optimized initial foreign body type recognition model; Obtain a foreign body type recognition test set; The optimized initial foreign body type recognition model is tested based on the foreign body type recognition test set. When the test result is qualified, the trained foreign body type recognition model is obtained.

7. The method for identifying foreign matter types for power channel safety monitoring according to claim 1, characterized in that: When it is determined that there is a type of foreign body with hidden dangers in the identification result, the risk caused by the foreign body is evaluated to obtain the target evaluation value, including: A Cartesian rectangular coordinate system is established according to the image to be identified when there is a foreign object type with hidden dangers in the identification result, and the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object are determined; Determine whether there is overlap between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object; If there is overlap between the power channel and the foreign object, the area of ​​the overlap region between the power channel and the foreign object is determined based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object; Obtaining a preset overlapping region judgment threshold corresponding to the foreign body type, and taking a ratio of the overlapping region area to the preset overlapping region judgment threshold as a first evaluation value; If there is no overlap between the power channel and the foreign object, determining the shortest distance between the power channel and the foreign object based on the coordinate information corresponding to the power channel and the coordinate information corresponding to the foreign object; Obtaining a preset distance judgment threshold corresponding to the foreign object type, and taking a ratio of the preset distance judgment threshold to the shortest distance as a second evaluation value; The first evaluation value and the second evaluation value are used as target evaluation values.

8. The method for identifying foreign matter types for power channel safety monitoring according to claim 7, characterized in that: When the target assessment value is determined to be greater than or equal to the preset risk assessment threshold, an early warning prompt is issued, including: If there is overlap between the power channel and the foreign object, the first evaluation value included in the target evaluation value is compared with the preset risk evaluation threshold, and when it is determined that the first evaluation value is greater than or equal to the preset risk evaluation threshold, an early warning prompt is issued; If there is no overlap between the power channel and the foreign object, the second evaluation value included in the target evaluation value is compared with the preset risk assessment threshold, and when it is determined that the second evaluation value is greater than or equal to the preset risk assessment threshold, an early warning prompt is issued.

9. The method for identifying foreign matter types for power channel safety monitoring according to claim 1, characterized in that: The types of foreign objects mentioned include surrounding illegal buildings, plastic greenhouses and temporary construction.

10. A foreign body type identification system for power channel safety monitoring, characterized in that: include: An acquisition module, used for acquiring an image to be identified for power channel safety monitoring; A preprocessing module, used for preprocessing the image to be identified; The recognition module is used to input the pre-processed image to be recognized into the pre-trained foreign body type recognition model for recognition and determine the recognition result; An evaluation module is used to evaluate the risk caused by the foreign matter when it is determined that there is a foreign matter type with hidden dangers in the identification result, and obtain a target evaluation value; The early warning module is used to issue an early warning prompt when it is determined that the target assessment value is greater than or equal to a preset risk assessment threshold.

Images