Substation camera adjusting method based on definition evaluation algorithm

Abstract

The invention provides a transformer substation camera adjusting method based on a definition evaluation algorithm, and relates to the technical field of transformer substation monitoring, and the method comprises the steps: adjusting a simulation camera according to the monitoring demand of a transformer substation; through rendering, testing, optimizing, adjusting, applying and verifying, the visual effect of the simulation camera is matched with that of the real camera; the camera parameter automatic adjusting system comprises an intelligent detecting and tracking module which is used for automatically detecting and tracking a target and adjusting focal length, angle and exposure parameters in real time through algorithm optimization so as to ensure that the target is always clear and visible; the behavior recognition and analysis module is used for recognizing and analyzing a behavior mode of the target by using a deep learning algorithm and automatically adjusting a monitoring strategy according to a recognition result; and the light adaptive module is used for automatically adjusting exposure, white balance and gain parameters according to the change of ambient light so as to ensure stable picture quality. According to the method, the imaging quality of the camera can be improved, and automatic focusing is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of substation monitoring, and particularly relates to a method for adjusting a substation camera based on a clarity evaluation algorithm. Background Art

[0002] Traditional power inspection methods mainly use a combination of manual methods and a small number of environmental monitors. However, the internal structure of substations is complex, there are many types of electrical equipment, and there are various abnormal situations. Moreover, when equipment malfunctions, it may pose risks to personal and equipment safety. In addition, manual inspections have problems such as high labor intensity and many subjective factors. With the continuous improvement of the intelligence level of power production, it is gradually developing towards intelligent substations and unmanned substations. Intelligent inspections based on deep learning already have applications using drones, inspection robots, and surveillance cameras as platforms.

[0003] In the prior art, since there are too many substation cameras for manual inspection in a timely manner, a substation intelligent inspection system is adopted. However, after a long time of use, substation cameras may become blurred and need to be adjusted to be clear. Currently, it is manually adjusted, and the cameras do not have a focusing function. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for adjusting a substation camera based on a clarity evaluation algorithm, aiming to solve the technical problems that after a long time of use, substation cameras may become blurred and need to be adjusted to be clear, currently it is manually adjusted, and the cameras do not have a focusing function. The specific technical solutions are as follows:

[0005] A method for adjusting a substation camera based on a clarity evaluation algorithm, the method comprising the following steps:

[0006] S100. Adjust the installation position, angle, and height of the simulation camera according to the monitoring requirements of the substation to ensure coverage of the target area, and adjust the focal length of the lens to obtain a clearer or wider picture;

[0007] S200. In the settings interface of the simulation camera, adjust the resolution, frame rate, and color settings to adapt to different monitoring environments and storage requirements, and turn on or off the motion detection function as needed;

[0008] S300. Match the clarity of the simulation camera in the substation three-dimensional model with the real camera, and obtain the detailed parameters of the real camera, specifically the key parameters of the focal length, aperture size, and viewing angle of the camera; then make corresponding settings for the simulation camera in the three-dimensional model to ensure that its parameters are consistent with the real camera; finally, through rendering, testing, optimization, adjustment, application, and verification, make the visual effect of the simulation camera match that of the real camera;

[0009] S400. Turn on the camera parameter automatic adjustment system, which includes: an intelligent detection and tracking module for automatically detecting and tracking moving objects or specific targets in the picture, and real-time adjusting the focal length, angle, and exposure parameters through algorithm optimization to ensure that the target is always clearly visible; a behavior recognition and analysis module for using deep learning algorithms to identify and analyze the behavior patterns of moving objects or specific targets, and automatically adjusting the monitoring strategy according to the recognition results; a light adaption module for automatically adjusting the exposure, white balance, and gain parameters according to the changes in environmental light to ensure stable picture quality.

[0010] Further, in step S300, obtaining the detailed parameters of the real camera specifically refers to the key parameters of the focal length, aperture size, and viewing angle of the camera; the specific steps for setting the simulation camera in the 3D model are as follows: find the setting option of the simulation camera in the 3D modeling software, and then adjust the focal length, aperture, and viewing angle of the simulation camera according to the parameters of the real camera to make it consistent with the real camera.

[0011] Further, in step S300, through rendering, testing, optimization, adjustment, application, and verification, make the visual effect of the simulation camera match that of the real camera. The specific steps are as follows: use a suitable renderer to render the 3D model to ensure that the imaging effect of the simulation camera is close to reality; observe the imaging effect of the simulation camera in the 3D model, compare it with the real camera, and according to the test results, fine-tune the parameters of the simulation camera to achieve the best matching effect; consider the light and shadow factors in the real environment and optimize the lighting and shadow effects in the 3D model; adjust the position and angle of the simulation camera to make it better integrate into the 3D model while maintaining the same visual effect as the real camera; apply the adjusted 3D model and simulation camera to the actual scene for verification, and ensure that the clarity, viewing angle, etc. of the simulation camera are consistent with the real camera to meet the actual application requirements.

[0012] Further, in the algorithm optimization of step S400, the focal length calculation formula is as follows:

[0013] f1 = wL / W

[0014] f2 = hL / H

[0015] Where f1 is the focal length of the lens in the horizontal direction, f2 is the focal length of the lens in the vertical direction, w is the width of the image, W is the width of the object being photographed, h is the height of the image, H is the height of the object being photographed, and L is the distance from the object being photographed to the lens;

[0016] The angle calculation formula is as follows:

[0017] β = 2 * tan(w / (2L))

[0018] q = 2 * tan(h / (2L))

[0019] Wherein, β is the horizontal field of view, q is the vertical field of view, w is the width of the image, h is the height of the image, and L is the distance from the object to be photographed to the lens;

[0020] The exposure time calculation formula is as follows:

[0021] τ = D1 / V1 = D2 / V2;

[0022] Wherein, τ is the exposure time, D1 is the pixel size, V1 is the moving speed of the image, D2 is the pixel accuracy, and V2 is the moving speed of the object.

[0023] Furthermore, the deep learning algorithm in the step S400 includes:

[0024] The forward propagation formula of the convolutional layer is as follows:

[0025]

[0026] Wherein, y(i, j) is the value at the position (i, j) in the output matrix, i and j are the row and column indices of the output, P and Q are the number of rows and columns of the kernel, that is, the height and width of the kernel, p and q are summation variables, traversing the row and column indices of the kernel, from 0 to P - 1 and Q - 1, x(i - p, j - q) is the value at the position (i - p, j - q) in the input matrix, k(p, q) is the weight of the convolution kernel, and b is the bias term;

[0027] The forward propagation formula of the pooling layer is as follows:

[0028] y(i, j) = max{x(p, q)|(p, q) ∈ R(i, j)}

[0029] Wherein, y(i, j) is the value of the output feature map at the position (i, j), and x(p, q) is the value of the input feature map within the pooling region R(i, j);

[0030] The forward propagation formula of the fully connected layer is as follows:

[0031] y = σ(Vx + b)

[0032] Wherein, y is the output vector, V is the weight matrix, x is the input vector, b is the bias term, and σ is the activation function;

[0033] The gradient descent algorithm formula is as follows:

[0034]

[0035] Where P t+1 is the point of the next iteration, (x t, y t ) is the current point, η is the learning rate, is the gradient of the loss function L at the point (x t , y t ).

[0036] Furthermore, in the step S400, the steps of automatically adjusting the exposure, white balance, and gain parameters according to the change of environmental light are as follows:

[0037] Enter the camera settings interface;

[0038] Select the exposure, white balance, and gain parameter adjustment options;

[0039] According to the change of environmental light, select a suitable automatic or manual adjustment mode;

[0040] If the manual adjustment mode is selected, adjust the ISO, shutter speed, aperture, and K value parameters according to actual needs;

[0041] Preview the adjustment effect to ensure that the picture brightness, color, and clarity meet the requirements;

[0042] Save the settings and exit the settings interface.

[0043] Furthermore, in the step S400, the camera parameter automatic adjustment system further includes: a weather adaptation module for automatically adjusting the image enhancement algorithm under rainy, snowy, or foggy weather conditions to reduce the impact of weather on the monitoring screen; a cloud intelligent management module for remotely managing and configuring various parameters of the camera by cloud platform users to achieve real-time monitoring and adjustment; a multi-camera cooperation module for multiple cameras to work together and achieve a wider monitoring coverage and more accurate target tracking through multiple camera cooperation algorithms.

[0044] Furthermore, the automatic adjustment of the image enhancement algorithm is based on the Retinex algorithm, and the image enhancement is achieved through the following steps:

[0045] Estimate the illumination image: L(x, y) is the illumination image of the input image I(x, y), (x, y) is the coordinate in the image, and logarithmic transformation is performed on the input image I(x, y) and the illumination image L(x, y) respectively to reduce the calculation amount. The calculation formula is as follows:

[0046] log(I(x, y)) = log(R(x, y)L(x, y)) = log(R(x, y)) + log(L(x, y))

[0047] where R(x, y) is the reflection image;

[0048] Reflection image solution: The solution of the reflection image is achieved through subtraction operation. The calculation formula is as follows:

[0049] Log(R(x,y)) = log(I(x,y)) - log(L(x,y));

[0050] Inverse logarithmic transformation: Perform an inverse logarithmic transformation on the result obtained from the subtraction to obtain the final result of Retinex algorithm image enhancement.

[0051] Furthermore, the algorithm for multiple cameras to work together is as follows:

[0052] w ij = f(D ij ) × (aP ij + bS ij )

[0053] where w ij represents the effectiveness of camera i for tracking target j, D ij represents the depth level of target j in camera i, f(D ij ) is a distance attenuation function or a spatial weight function, P ij represents the priority of target j, which is determined by the category, feature area, behavior type, or motion trajectory characteristics of target j observed by camera i, S ij represents the size of the segmented image of target j in camera i, and a and b are weight coefficients.

[0054] Furthermore, after step S400, the method further includes the following steps:

[0055] S500. Turn on or off the privacy protection function as needed to protect the privacy of specific areas;

[0056] S600. Finally, set the network connection parameters of the camera to ensure that the camera can be normally connected to the network for data transmission and control.

[0057] A substation camera adjustment method based on a clarity evaluation algorithm provided by the present invention has the following beneficial effects:

[0058] (1). The present invention can automatically identify the inspection scenario and light conditions, and automatically adjust the shooting parameters of the camera according to this information, ensuring that the camera can maintain the best shooting effect in different environments, reducing the need for manual intervention, and improving the imaging quality of the camera;

[0059] (2). The present invention utilizes the three-dimensional model inherent in the substation to match the clarity of the simulated camera with that of the real camera, thereby improving the configuration efficiency and accuracy of the camera;

[0060] (3) By adjusting parameters such as the resolution, frame rate, and focal length of the camera, the present invention can obtain a clearer and smoother picture, thereby improving the detection accuracy and efficiency. As a result, more details can be captured, false alarms and missed detections can be reduced, and the overall detection effect can be enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a schematic flowchart of a method for adjusting a substation camera based on a sharpness evaluation algorithm provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings provided by the present invention. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the drawings are all in a very simplified form and use non-precise scales, only for the purpose of facilitating and clearly assisting in explaining the objectives of the embodiments of the present invention.

[0063] Embodiment 1

[0064] This embodiment provides a method for adjusting a substation camera based on a sharpness evaluation algorithm. Referring to Figure 1 as shown, the method includes the following steps:

[0065] S100. Adjust the installation position, angle, and height of the simulation camera according to the monitoring requirements of the substation to ensure coverage of the target area, and adjust the focal length of the lens to obtain a clearer or wider picture.

[0066] S200. In the settings interface of the simulation camera, adjust the resolution, frame rate, and color settings to adapt to different monitoring environments and storage requirements, and turn on or off the motion detection function as needed.

[0067] S300. Match the sharpness of the simulation camera in the 3D model of the substation with that of the real camera, and obtain the detailed parameters of the real camera, specifically the key parameters of the focal length, aperture size, and viewing angle of the camera; then make corresponding settings for the simulation camera in the 3D model to ensure that its parameters are consistent with those of the real camera; finally, through rendering, testing, optimization, adjustment, application, and verification, make the visual effect of the simulation camera match that of the real camera.

[0068] In one embodiment, obtaining the detailed parameters of the real camera specifically refers to the key parameters of the focal length, aperture size, and viewing angle of the camera; the specific steps for making corresponding settings for the simulation camera in the 3D model are as follows: find the settings option of the simulation camera in the 3D modeling software, and then adjust the focal length, aperture, and viewing angle of the simulation camera according to the parameters of the real camera to make them consistent with the real camera.

[0069] In one embodiment, through rendering, testing, optimization, adjustment, application, and verification, the visual effect of the simulated camera is made to match that of the real camera. The specific steps are as follows: Use a suitable renderer to render the 3D model to ensure that the imaging effect of the simulated camera is close to that of the real one; Observe the imaging effect of the simulated camera in the 3D model, compare it with the real camera, and according to the test results, fine-tune the parameters of the simulated camera to achieve the best matching effect; Consider the light and shadow factors in the real environment and optimize the lighting and shadow effects in the 3D model; Adjust the position and angle of the simulated camera so that it better integrates into the 3D model while maintaining the same visual effect as the real camera; Apply the adjusted 3D model and the simulated camera to the actual scenario for verification, and ensure that the clarity, viewing angle, etc. of the simulated camera are consistent with those of the real camera to meet the actual application requirements.

[0070] S400. Turn on the automatic camera parameter adjustment system. The automatic camera parameter adjustment system includes: an intelligent detection and tracking module for automatically detecting and tracking moving objects or specific targets in the picture, such as a human figure, and optimizing the algorithm to adjust the focal length, angle, and exposure parameters in real time to ensure that the target is always clearly visible; a behavior recognition and analysis module for using deep learning algorithms to identify and analyze the behavior patterns of moving objects or specific targets, such as wandering or detecting left-behind objects, and automatically adjusting the monitoring strategy according to the recognition results, such as increasing the video recording density or triggering an alarm; a light adaptive module for automatically adjusting the exposure, white balance, and gain parameters according to the changes in the ambient light to ensure stable picture quality.

[0071] In one embodiment, in the algorithm optimization, the focal length (f) is related to the field of view (FOV), the width (W) of the object, the height (H) of the object, and the distance (L) from the object to the lens. The focal length calculation formula is as follows:

[0072] f1 = wL / W

[0073] f2 = hL / H

[0074] Where f1 is the focal length of the lens in the horizontal direction, f2 is the focal length of the lens in the vertical direction, w is the width of the image (the imaging width of the object on the ccd target surface), W is the width of the object being photographed, h is the height of the image (the imaging height of the object on the ccd target surface), H is the height of the object being photographed, and L is the distance from the object being photographed to the lens;

[0075] The angle refers to the field of view angle, including the horizontal field of view angle (β) and the vertical field of view angle (q). The angle calculation formula is as follows:

[0076] β = 2 * tan(w / (2L))

[0077] q = 2 * tan(h / (2L))

[0078] Among them, β is the horizontal field of view angle, q is the vertical field of view angle, w is the width of the image, h is the height of the image, and L is the distance from the object to be photographed to the lens;

[0079] The exposure time is the time for the shutter to open in order to project light onto the photosensitive surface of the photographic light-sensitive material. The exposure time is inversely proportional to the actual frame rate. The exposure time calculation formula is as follows:

[0080] τ = D1 / V1 = D2 / V2;

[0081] Among them, τ is the exposure time, D1 is the pixel size, V1 is the movement speed of the image, D2 is the pixel accuracy, and V2 is the movement speed of the object;

[0082] For example, if the movement speed of the object is 150 mm / s and it moves along the horizontal direction of the chip, the camera is a 1 / 2 chip (6.4 mm * 4.8 mm), the horizontal field of view length is 20 mm, and the pixel size is 4.65 um, then the exposure time can be calculated as 97 us.

[0083] In one embodiment, the deep learning algorithm includes:

[0084] The forward propagation formula of the convolutional layer is as follows:

[0085]

[0086] Among them, y(i,j) is the value at the position (i,j) in the output matrix, i and j are the row and column indices of the output, P and Q are the number of rows and columns of the kernel, that is, the height and width of the kernel, p and q are the summation variables, traversing the row and column indices of the kernel from 0 to P-1 and Q-1, x(i-p,j-q) is the value at the position (i-p,j-q) in the input matrix, k(p,q) is the weight of the convolutional kernel, and b is the bias term;

[0087] The pooling layer is usually used to reduce the dimension of the feature map and reduce the amount of calculation. The forward propagation formula of the pooling layer is as follows:

[0088] y(i,j) = max{x(p,q)|(p,q) ∈ R(i,j)}

[0089] Among them, y(i,j) is the value of the output feature map at the position (i,j), and x(p,q) is the value of the input feature map within the pooling region R(i,j);

[0090] The fully connected layer flattens the feature map of the previous layer into a vector and applies a linear transformation and an activation function to generate the output. The forward propagation formula of the fully connected layer is as follows:

[0091] y = σ(Vx + b)

[0092] Among them, y is the output vector, V is the weight matrix, x is the input vector, b is the bias term, and σ is the activation function;

[0093] In deep learning, the gradient descent algorithm is used to optimize the parameters of the model to minimize the loss function. In each iteration, the model parameters are updated by subtracting the value of the learning rate multiplied by the gradient from the previous step size. The formula for the gradient descent algorithm is as follows:

[0094]

[0095] Among them, P t+1 is the point for the next iteration, (x t , y t ) is the current point, η is the learning rate, is the gradient of the loss function L at the point (x t , y t ).

[0096] In one embodiment, the steps of automatically adjusting the exposure, white balance, and gain parameters according to the change of ambient light are as follows:

[0097] Enter the settings interface of the camera;

[0098] Select the exposure, white balance, and gain parameter adjustment options;

[0099] According to the change of ambient light, select the appropriate automatic or manual adjustment mode;

[0100] If the manual adjustment mode is selected, adjust the ISO, shutter speed, aperture, and K value parameters according to actual needs;

[0101] Preview the adjustment effect to ensure that the picture brightness, color, and clarity meet the requirements;

[0102] Save the settings and exit the settings interface.

[0103] Specifically, the exposure adjustment includes:

[0104] Automatic exposure: The camera has an automatic exposure function, which automatically adjusts the exposure parameters according to the ambient light to ensure appropriate picture brightness. The user only needs to set the camera to the automatic exposure mode;

[0105] Manual exposure: For scenes that require precise control of exposure, the user selects the manual exposure mode and manually adjusts the ISO, shutter speed, and aperture parameters;

[0106] The white balance adjustment includes:

[0107] Auto White Balance: The camera automatically adjusts the white balance according to the ambient light to restore the true colors. When the user selects the auto white balance mode, the camera will automatically adapt to the color temperature of different light sources.

[0108] Manual White Balance: If more precise color control is needed, the user can select the manual white balance mode and choose the corresponding preset mode (such as sunlight, overcast, fluorescent light, etc.) according to the shooting environment or manually set the K value.

[0109] Custom White Balance: Calibrate the color temperature by shooting a white object.

[0110] Gain Adjustment: Combined with exposure adjustment, it is used to increase the brightness of the picture in a dim environment. The user appropriately adjusts the gain value according to the change of the ambient light to ensure the balance of picture clarity and brightness.

[0111] In one embodiment, the camera parameter automatic adjustment system further includes: a weather adaptation module, which is used to automatically adjust the image enhancement algorithm under rainy, snowy or foggy weather conditions to reduce the impact of the weather on the monitoring picture; a cloud intelligent management module, which is used for users to remotely manage and configure various parameters of the camera through the cloud platform to achieve real-time monitoring and adjustment; a multi-camera cooperation module, which is used for multiple cameras to work together to achieve a wider monitoring coverage and more accurate target tracking through the multi-camera cooperation algorithm.

[0112] In a preferred embodiment, the automatic adjustment of the image enhancement algorithm is based on the Retinex algorithm, and the image enhancement is achieved through the following steps:

[0113] Estimate the illumination image: Let L(x,y) be the illumination image of the input image I(x,y), and (x,y) be the coordinates in the image. The logarithmic transformation is performed on the input image I(x,y) and the illumination image L(x,y) respectively to reduce the calculation amount. The calculation formula is as follows:

[0114] log(I(x,y)) = log(R(x,y)L(x,y)) = log(R(x,y)) + log(L(x,y))

[0115] where R(x,y) is the reflection image;

[0116] Solve the reflection image: The reflection image is solved through subtraction operation. The calculation formula is as follows:

[0117] Log(R(x,y)) = log(I(x,y)) - log(L(x,y));

[0118] Inverse logarithmic transformation: Perform the inverse logarithmic transformation on the result obtained by subtraction to obtain the final result of the Retinex algorithm image enhancement.

[0119] In addition, there is also a defogging algorithm based on the dark channel prior, and the calculation formula of its transmittance is as follows:

[0120] t(x) = 1 - ω * min(c∈{r,g,b}) * min((y∣y∈Ω(x)) * min(Ic(y) / Ac))

[0121] Where t(x) is the transmittance, ω is the adjustment factor, min(c∈{r,g,b}) is the minimum value obtained by separately calculating the RGB three channels, Ω(x) is the local area centered on pixel x (such as a 15×15 window), Ic(y) is the pixel value of the input image at position y and channel c, and Ac is the global atmospheric light value.

[0122] In a preferred embodiment, the algorithm for multiple cameras to work together is as follows:

[0123] w ij = f(D ij ) × (aP ij + bS ij )

[0124] Where w ij represents the effectiveness of camera i for tracking target j, D ij represents the depth level of target j in camera i, f(D ij ) is a distance attenuation function or a spatial weight function, P ij represents the priority of target j, which is determined by the category, feature area, behavior type or motion trajectory characteristics of target j observed by camera i, S ij represents the size of the segmented image of target j in camera i, and a and b are weight coefficients.

[0125] In one embodiment, after step S400, the method further includes the following steps:

[0126] S500. Turn on or off the privacy protection function as needed to protect the privacy of specific areas;

[0127] S600. Finally, set the network connection parameters of the camera to ensure that the camera can be normally connected to the network for data transmission and control.

[0128] Those skilled in the art of this technology should understand that the present invention can be implemented in many other specific forms without departing from the spirit and scope of the present invention. Based on the embodiments of the present invention, any changes and modifications made by those of ordinary skill in the art of the present invention according to the above disclosure fall within the protection scope of the claims.

Claims

1. A substation camera adjustment method based on a clarity evaluation algorithm, characterized in that: The method comprises the following steps: S100. Adjust the installation position, angle and height of the simulation camera according to the monitoring requirements of the substation to ensure coverage of the target area, and adjust the focal length of the lens to obtain a clearer or wider picture; S200. In the setting interface of the simulation camera, adjust the resolution, frame rate and color settings to adapt to different monitoring environments and storage requirements, and turn on or off the motion detection function as needed; S300, matching the definition of the simulated camera in the three-dimensional model of the substation with the real camera, and obtaining detailed parameters of the real camera, specifically the focal length, aperture size and viewing angle key parameters of the camera; then setting the simulated camera accordingly in the three-dimensional model to ensure that its parameters are consistent with those of the real camera; finally, through rendering, testing, optimization, adjustment, application and verification, the visual effect of the simulated camera is matched with that of the real camera; S400. Turn on the camera parameter automatic adjustment system. The camera parameter automatic adjustment system includes: an intelligent detection and tracking module, which is used to automatically detect and track moving objects or specific targets in the picture, and adjust the focal length, angle and exposure parameters in real time through algorithm optimization to ensure that the target is always clearly visible; a behavior recognition and analysis module, which is used to use deep learning algorithms to identify and analyze the behavior patterns of moving objects or specific targets, and automatically adjust the monitoring strategy according to the recognition results; a light adaptation module, which is used to automatically adjust the exposure, white balance and gain parameters according to changes in ambient light to ensure stable picture quality.

2. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: In the step S300, the detailed parameters of the real camera are obtained, specifically the focal length, aperture size and viewing angle key parameters of the camera; the specific steps of making corresponding settings for the simulated camera in the three-dimensional model are: finding the setting options of the simulated camera in the three-dimensional modeling software, and then adjusting the focal length, aperture and viewing angle of the simulated camera according to the parameters of the real camera to make them consistent with the real camera.

3. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 2 is characterized in that: In the step S300, the visual effect of the simulated camera is matched with the real camera through rendering, testing, optimization, adjustment, application and verification. The specific steps are: using a suitable renderer to render the three-dimensional model to ensure that the imaging effect of the simulated camera is close to reality; Observe the imaging effect of the simulated camera in the 3D model, compare it with the real camera, and fine-tune the parameters of the simulated camera based on the test results to achieve the best matching effect; consider the light and shadow factors in the real environment to optimize the lighting and shadow effects in the 3D model; adjust the position and angle of the simulated camera to better integrate it into the 3D model while maintaining the same visual effect as the real camera; Apply the adjusted 3D model and simulated camera to the actual scene for verification, and ensure that the clarity, viewing angle, etc. of the simulated camera are consistent with the real camera to meet actual application requirements.

4. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: In the algorithm optimization of step S400, the focal length calculation formula is as follows: f1=wL / W f2=hL / H Where, f1 is the horizontal focal length of the lens, f2 is the vertical focal length of the lens, w is the width of the image, W is the width of the object, h is the image height, H is the height of the object, and L is the distance from the object to the lens; The angle calculation formula is as follows: β=2*tan(w / (2L)) q=2*tan(h / (2L)) Where, β is the horizontal field of view, q is the vertical field of view, w is the width of the image, h is the height of the image, and L is the distance from the object to the lens; The exposure time calculation formula is as follows: τ=D1 / V1=D2 / V2; Among them, τ is the exposure time, D1 is the pixel size, V1 is the image movement speed, D2 is the pixel accuracy, and V2 is the object movement speed.

5. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: The deep learning algorithm of step S400 includes: The forward propagation formula of the convolutional layer is as follows: Where y(i,j) is the value at position (i,j) in the output matrix, i and j are the row and column indices of the output, P and Q are the number of rows and columns of the kernel, that is, the height and width of the kernel, p and q are summation variables, traversing the row and column indices of the kernel, from 0 to P-1 and Q-1,, x(ip,jq) is the value at position (ip,jq) in the input matrix, k(p,q) is the weight of the convolution kernel, and b is the bias term; The forward propagation formula of the pooling layer is as follows: y(i,j)=max{x(p,q)|(p,q)∈R(i,j)} Among them, y(i,j) is the value of the output feature map at position (i,j), and x(p,q) is the value of the input feature map in the pooling area R(i,j); The forward propagation formula of the fully connected layer is as follows: y=σ(Vx+b) Where y is the output vector, V is the weight matrix, x is the input vector, b is the bias term, and σ is the activation function; The gradient descent algorithm formula is as follows: Among them, P t+1 is the point of the next iteration, (x t ,y t ) is the current point, η is the learning rate, is the loss function L at point (x t ,y t )’s gradient.

6. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: In step S400, the steps of automatically adjusting exposure, white balance and gain parameters according to changes in ambient light are: Enter the camera settings interface; Select exposure, white balance and gain parameter adjustment options; Select the appropriate automatic or manual adjustment mode according to changes in ambient light; If you choose manual adjustment mode, adjust ISO, shutter speed, aperture, and K value parameters according to actual needs; Preview the adjustment effect to ensure that the image brightness, color and clarity meet the requirements; Save the settings and exit the settings interface.

7. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: In step S400, the camera parameter automatic adjustment system also includes: a weather adaptation module, which is used to automatically adjust the image enhancement algorithm under rainy, snowy or foggy weather conditions to reduce the impact of weather on the monitoring screen; a cloud-based intelligent management module, which is used to remotely manage and configure various parameters of the camera through cloud platform users to achieve real-time monitoring and adjustment; a multi-camera collaborative module, which is used for multiple cameras to work together, and achieves wider monitoring coverage and more accurate target tracking through multiple camera collaborative working algorithms.

8. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 7 is characterized in that: The automatic adjustment image enhancement algorithm is based on the Retinex algorithm and achieves image enhancement through the following steps: Estimated illumination image: L(x,y) is the illumination image of the input image I(x,y), (x,y) is the coordinate in the image, and the input image I(x,y) and the illumination image L(x,y) are logarithmically transformed to reduce the amount of calculation. The calculation formula is as follows: log(I(x,y))=log(R(x,y)L(x,y))=log(R(x,y))+log(L(x,y)) Among them, R(x,y) is the reflected image; Reflection image solution: The reflection image is solved by subtraction operation. The calculation formula is as follows: Log(R(x,y))=log(I(x,y))-log(L(x,y)); Inverse logarithmic transformation: Perform an inverse logarithmic transformation on the result obtained by subtraction to obtain the final result of the Retinex algorithm image enhancement.

9. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 7 is characterized in that: The collaborative working algorithm of multiple cameras is: w ij =f(D ij )×(aP ij +bS ij ) Among them, w ij represents the effectiveness of camera i for tracking target j, D ij represents the depth level of target j in camera i, f(D ij ) is the distance decay function or spatial weight function, P ij represents the priority of target j, which is determined by the category, feature area, behavior type or motion trajectory characteristics of target j observed by camera i. ij represents the segmented image size of target j in camera i, and a and b are weight coefficients.

10. The substation camera adjustment method based on the clarity evaluation algorithm according to claim 1 is characterized in that: The method further comprises the following steps after step S400: S500, turning on or off the privacy protection function as needed to protect the privacy of a specific area; S600. Finally, set the network connection parameters of the camera to ensure that the camera can be normally connected to the network for data transmission and control.

Images