CN102955940B - A kind of transmission line of electricity object detecting system and method - Google Patents

Abstract

The invention discloses a power transmission line object detection system, which includes a camera, the camera is connected to a video server through RS485, the video output by the video server is sent to an image acquisition board, the image acquisition board is connected to a host computer, and the host computer is connected to a base station. The computer is connected with the alarm, and the image acquisition board is also connected with the DSP high-speed compression circuit and the memory. The image acquisition board receives the control command sent by the host computer and uploads it to the video server. The video server controls the camera to take pictures. The picture is uploaded to the image acquisition board through the video server, and the image in the image acquisition board is registered by the DSP high-speed compression circuit with the image compression and the reference image in the memory, and an object detection method is disclosed. The invention can detect intrusion and Foreign objects are detected, and if abnormal objects are found, an alarm will be issued, so that the staff can take effective measures to avoid transmission line accidents, and the alarm information will be uploaded step by step.

Description

A transmission line object detection system and method

technical field

The invention relates to an object detection system and method, in particular to a power transmission line object detection system and method.

Background technique

With the rapid development of power construction and the continuous expansion of power grid scale, more and more power grids are built under complex terrain conditions, and transmission lines are characterized by large dispersion, long distances, and difficulty in maintenance. The power sector currently adopts regular manual inspections or helicopter inspections. The general inspection cycle is one month, and the environmental conditions of the line and its surroundings are unknown during the inspection period, which has laid a huge burden on the safe operation of the transmission line. Hidden danger.

At present, most remote monitoring systems only realize the monitoring of transmission lines, and there are no mature technologies for intrusion detection and foreign object detection.

Contents of the invention

The object of the present invention is to solve the above-mentioned problems and provide a method for detecting objects on transmission lines, which has the advantage of automatically identifying objects beside the transmission lines.

In order to achieve the above object, the present invention adopts the following technical solutions:

A power transmission line object detection system, which includes a camera, the camera is connected to the video server through RS485, the video output by the video server is sent to the image acquisition board, the image acquisition board is connected to the upper computer, the upper computer is connected to the base station, and the upper computer is connected to the alarm connected, the image acquisition board is also connected with the DSP high-speed compression circuit and memory, the image acquisition board receives the control command sent by the host computer, uploads to the video server, the video server controls the camera to take pictures, and the pictures taken by the camera pass through the video server Upload to the image acquisition board, and the image in the image acquisition board is registered by the DSP high-speed compression circuit to compress the image and the reference image in the memory.

A transmission line object detection method based on a transmission line object detection system, the specific steps are:

Step 1: read the original background image, and use the adaptive background mixed Gaussian model modeling method to create an initial mixed Gaussian background model;

Step 2: capture the video frame, detect the matching between the new pixel value and the template, if there is a partial match, modify the Gaussian parameters, and if there is no match, build another mixed Gaussian background model;

Step 3: After building another mixed Gaussian background model, return to Step 2;

Step 4: Generate a background image;

Step 5: Read the current video frame, and use the background difference method to extract the moving target in the current frame;

Step 6: Use the Kalman filter algorithm to accurately extract the moving targets within the designated monitoring area;

Step 7: Then use the characteristics of each object to establish a model of each object to identify the type of moving object within the monitoring range through the method of support vector machine;

Step 8: If it is detected that large-scale machinery construction activities enter the monitoring area, the alarm will be activated; when someone is found to enter the monitoring area, the person will be further identified and tracked, and determined according to their movement time within the monitoring area Its motion behavior is to judge whether there is an act of digging the optical cable privately, and if so, the upper computer controls the alarm to alarm.

The modeling principle of the mixed Gaussian background model in the first step is:

Assume that the gray value of each pixel is described by K Gaussian distributions. Usually the K value is 3 to 5. The size of the K value depends on the computer memory and the speed requirements of the algorithm. The larger the K value, the better the ability to deal with gray scale changes. Stronger, the correspondingly longer processing time is required. Define the pixel gray value to be represented by the variable X _t , and its probability density function is represented by K three-dimensional Gaussian functions:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; \&eta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>$

where ω _i,t is the weight of the i-th Gaussian distribution at time t, and η(X _t , μ _{i, t} , ∑ _{i, t} ) is the i-th Gaussian distribution at time t, with mean value μ _{i, t} and covariance ∑ _{i, t} ;

$\mathrm{<span\; class="notranslate">\; \&eta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; )</span>}^{\frac{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; T</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>}$

In the formula, i=1,...,K; n represents the dimension of X _t , assuming R, G, B, the three channels are independent of each other and have the same variance, then we have ${\mathrm{\&Sigma;}}_{i,t}={\mathrm{\&Sigma;}}_{i,t}{\mathrm{\&sigma;}}_i^{2}I,$ Represents the variance, and I represents the identity matrix.

After the background image is generated, it is necessary to extract the moving target in the current frame. The principle of the background difference method is as follows: first, set B _k as the background image, f _k as the current frame image, and the difference image as D _k , then D _k (x , y)=|f _x (x, y)-B _k-1 (x, y)|, (x, y) is the pixel position, let R _k be the binary image after difference, and perform connectivity analysis on R _k , when the area of a certain connected region is greater than a certain threshold, it is considered that the detected target appears, and this connected region is considered to be the detected target image.

T is the set binarization threshold.

The concrete steps of described step six are:

(6-1) Establish the linear stochastic differential equation of the Kalman filter system: X(k)=A X(k-1)+B U(k)+W(k), the measured value of the Kalman filter system : Z(k)=H X(k)+V(k); X(k) is the system state at time k, U(k) is the control quantity of the system at time k, A and B are system parameters, Z (k) is the measured value at time k, H is the parameter of the measurement system, W(k) and V(k) represent the process and measurement noise respectively;

(6-2) Assuming that the current system state is k, according to the linear stochastic differential equation of the Kalman filter system, use the state of the previous frame to predict the state of the current frame:

X(k|k-1)=A·X(k-1|k-1)+B·U(k)(1)

Among them, X(k|k-1) is the result of using the state prediction of the previous frame, X(k-1|k-1) is the result of the optimal state of the previous frame, and U(k) is the control of the current frame state amount, if there is no control amount, it is 0;

(6-3) Use formula (2) to update the covariance of X(k|k-1),

P(k|k-1)=A·P(k-1|k-1)· ^AT +Q(2), P(k|k-1) is X(k|k-1)

The corresponding covariance, P(k-1|k-1) is the covariance corresponding to X(k-1|k-1), ^AT represents the transpose matrix of A, and Q is the covariance of the system process;

(6-4) Collect the measured value of the current frame state, and combine the predicted value and measured value of the frame state to obtain the optimal estimated value X(k|k) of the current frame state:

X(k|k)=X(k|k-1)+Kg(k)·(Z(k)-H·X(k|k-1))(3)

Among them, Kg is Kalman Gain (KalmanGain); Kg(k)=P(k|k-1)H ^T /(H·P(k|k-1)· ^HT +R)(4);

(6-5) Constantly update the covariance of K state X(k|k): P(k|k)=(I-Kg(k)·H)P(k|k-1)(5)

Where I is a matrix of 1, for a single model and single measurement, I=1; when the system enters the k+1 state, P(k|k) is P(k|k-1) of the formula (2).

In the demarcated monitoring area, through background modeling, difference processing, morphological operations, etc. to extract more accurate moving targets, and then use the characteristics of cars, people and animals through the method of support vector machine to establish the vehicle, people and The animal model is used to identify the types of moving objects within the monitoring range. The detected moving foreground targets may include tall construction machinery, moving human objects, and wild animals. Generally, the color characteristics of tall construction machinery are more obvious and their area is larger. By calculating the foreground The area of the target area and the aspect ratio of the circumscribed rectangle are used to distinguish the foreground target. If it is detected that construction activities of large machinery enter the monitoring area, the front-end alarm device will be activated. When it is found that someone enters the monitoring range, the above methods are used to further identify and track the person, and determine their movement behavior according to their movement time within the monitoring range to see if there is any behavior such as privately digging optical cables.

Beneficial effects of the present invention: the present invention can be used to detect intrusion and foreign objects, and if an abnormal object is found, an alarm is issued, so that staff can take effective measures to avoid transmission line accidents. Remind the human target to steal national power equipment is an illegal act, and upload the alarm information level by level.

Description of drawings

Fig. 1 is a system block diagram of the present invention;

Fig. 2 is a partial flow chart of the present invention.

Among them, 1. Image acquisition board, 2. Host computer, 3. Base station, 4. Camera, 5. Video server, 6. DSP high-speed compression circuit, 7. Memory, 8. Alarm.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, a transmission line object detection system includes a camera 4, the camera 4 is connected to the video server 5 through RS485, the video output by the video server 5 is sent to the image acquisition board 1, and the image acquisition board 1 is connected to the upper computer 2, The upper computer 2 is connected with the base station 3, the upper computer 2 is connected with the alarm 8, and the image acquisition board 1 is also connected with the DSP high-speed compression circuit 6 and the memory 7, and the image acquisition board 1 receives the control command sent by the upper computer 2 , upload to the video server 5, the video server 5 controls the camera 4 to take pictures, the picture taken by the camera 4 is uploaded to the image acquisition board 1 through the video server 5, and the image in the image acquisition board 1 is compressed by the DSP high-speed compression circuit 6 and stored in the memory 7 for registration.

As shown in Figure 2, a transmission line object detection method based on the transmission line object detection system, the specific steps are:

Step 1: read the original background image, and use the adaptive background mixed Gaussian model modeling method to create an initial mixed Gaussian background model;

Step 2: capture the video frame, detect the matching between the new pixel value and the template, if there is a partial match, modify the Gaussian parameters, and if there is no match, build another mixed Gaussian background model;

Step 3: After building another mixed Gaussian background model, return to Step 2;

Step 4: Generate a background image;

Step 5: Read the current video frame, and use the background difference method to extract the moving target in the current frame;

Step 6: Use the Kalman filter algorithm to accurately extract the moving targets within the designated monitoring area;

Step 7: Then use the characteristics of each object to establish a model of each object to identify the type of moving object within the monitoring range through the method of support vector machine;

Step 8: If it is detected that large-scale machinery construction activities enter the monitoring area, the alarm will be activated; when someone is found to enter the monitoring area, the person will be further identified and tracked, and determined according to their movement time within the monitoring area Its motion behavior is to judge whether there is an act of digging the optical cable privately, and if so, the upper computer controls the alarm to alarm.

The modeling principle of the mixed Gaussian background model in the first step is:

Assume that the gray value of each pixel is described by K Gaussian distributions. Usually, the value of K is 3 to 5. The value of K depends on the computer memory and the speed requirements of the algorithm. The larger the value of K, the better the ability to deal with gray scale changes. Stronger, the correspondingly longer processing time is required. Define the pixel gray value to be represented by the variable X _t , and its probability density function is represented by K three-dimensional Gaussian functions:

$\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&eta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>$

where ω _i,t is the weight of the i-th Gaussian distribution at time t, and η(X _t , μ _{i, t} , ∑ _{i, t} ) is the i-th Gaussian distribution at time t, with mean value μ _{i, t} and covariance ∑ _{i, t} ;

$\mathrm{<span\; class="notranslate">\; \&eta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; )</span>}^{\frac{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; T</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; )</span>}$

In the formula, i=1,...,K; n represents the dimension of X _t , assuming R, G, B, the three channels are independent of each other and have the same variance, then we have ${\mathrm{\&Sigma;}}_{i,t}={\mathrm{\&Sigma;}}_{i,t}{\mathrm{\&sigma;}}_i^{2}I,$ Represents the variance, and I represents the identity matrix.

After the background image is generated, it is necessary to extract the moving target in the current frame. The principle of the background difference method is as follows: first, set B _k as the background image, f _k as the current frame image, and the difference image as D _k , then D _k (x , y)=|f _x (x, y)-B _k-1 (x, y)|, (x, y) is the pixel position, let R _k be the binary image after difference, and perform connectivity analysis on R _k , when the area of a certain connected region is greater than a certain threshold, it is considered that the detected target appears, and this connected region is considered to be the detected target image.

T is the set binarization threshold.

The concrete steps of described step six are:

(6-1) Establish the linear stochastic differential equation of the Kalman filter system: X(k)=A X(k-1)+B U(k)+W(k), the measured value of the Kalman filter system : Z(k)=H X(k)+V(k); X(k) is the system state at time k, U(k) is the control quantity of the system at time k, A and B are system parameters, Z (k) is the measured value at time k, H is the parameter of the measurement system, W(k) and V(k) represent the process and measurement noise respectively;

(6-2) Assuming that the current system state is k, according to the linear stochastic differential equation of the Kalman filter system, use the state of the previous frame to predict the state of the current frame:

X(k|k-1)=A·X(k-1|k-1)+B·U(k)(1)

Among them, X(k|k-1) is the result of using the state prediction of the previous frame, X(k-1|k-1) is the result of the optimal state of the previous frame, and U(k) is the control of the current frame state amount, if there is no control amount, it is 0;

(6-3) Use formula (2) to update the covariance of X(k|k-1),

P(k|k-1)=A·P(k-1|k-1)· ^AT +Q(2), P(k|k-1) is X(k|k-1)

The corresponding covariance, P(k-1|k-1) is the covariance corresponding to X(k-1|k-1), ^AT represents the transpose matrix of A, and Q is the covariance of the system process;

(6-4) Collect the measured value of the current frame state, and combine the predicted value and measured value of the frame state to obtain the optimal estimated value X(k|k) of the current frame state:

X(k|k)=X(k|k-1)+Kg(k)·(Z(k)-H·X(k|k-1))(3)

Among them, Kg is Kalman Gain (KalmanGain); Kg(k)=P(k|k-1)H ^T /(H·P(k|k-1)· ^HT +R)(4);

(6-5) Constantly update the covariance of K state X(k|k): P(k|k)=(I-Kg(k)·H)P(k|k-1)(5)

Where I is a matrix of 1, for a single model and single measurement, I=1; when the system enters the k+1 state, P(k|k) is P(k|k-1) of the formula (2).

In the demarcated monitoring area, through background modeling, difference processing, morphological operations, etc. to extract more accurate moving targets, and then use the characteristics of cars, people and animals through the method of support vector machine to establish the vehicle, people and The animal model is used to identify the types of moving objects within the monitoring range. The detected moving foreground targets may include tall construction machinery, moving human objects, and wild animals. Generally, the color characteristics of tall construction machinery are more obvious and their area is larger. By calculating the foreground The area of the target area and the aspect ratio of the circumscribed rectangle are used to distinguish the foreground target. If it is detected that construction activities of large machinery enter the monitoring area, the front-end alarm device will be activated. When it is found that someone enters the monitoring range, the above methods are used to further identify and track the person, and determine their movement behavior according to their movement time within the monitoring range to see if there is any behavior such as privately digging optical cables.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (3)

1. A power transmission line object detection method is characterized in that the image acquisition board is uploaded to the video server after receiving the control command sent by the host computer, and the video server controls the camera to take pictures, and the picture taken by the camera is uploaded to the image through the video server Acquisition board, the image in the image acquisition board is registered by the DSP high-speed compression circuit to compress the image and the reference image in the memory, and the specific steps are: Step 1: Read the original background image and generate the background; Step 2: read the current video frame, and use the background difference method to extract the moving target in the current frame; Step 3: Use the Kalman filter algorithm to accurately extract the moving targets within the designated monitoring area; Step 4: Then use the characteristics of each object to establish a model of each object to identify the type of moving object within the monitoring range through the method of support vector machine; Step 5: If it is detected that there are large-scale machinery construction activities entering the monitoring area, the alarm will be activated; when someone is found to enter the monitoring area, the person will be further identified and tracked, and determined according to their movement time within the monitoring area Its motion behavior is to judge whether there is a behavior of digging the optical cable privately, and if so, the host computer will control the alarm to alarm; The concrete steps of described step one are: (1-1) Read the original background image and create an initial mixed Gaussian background model; (1-2) Capture the video frame, detect the matching between the new pixel value and the template, if it partially matches, then modify the Gaussian parameters, and if it does not match, then build a new model; (1-3) Obtain a new mixed Gaussian background model, then return to step (1-2); (1-4) Generate background. 2. A kind of transmission line object detection method as claimed in claim 1, is characterized in that, the specific steps of described step 3 are: (6-1) Establish the linear stochastic differential equation of the Kalman filter system: X(k)=A X(k-1)+B U(k)+W(k), the measured value of the Kalman filter system : Z(k)=H X(k)+V(k); X(k) is the system state at time k, U(k) is the control quantity of the system at time k, A and B are system parameters, Z (k) is the measured value at time k, H is the parameter of the measurement system, W(k) and V(k) represent the process and measurement noise respectively; (6-2) Assuming that the current system state is k, according to the linear stochastic differential equation of the Kalman filter system, the state of the current frame is predicted by using the state of the previous frame: X(k|k-1)＝A·X(k-1|k-1)+B·U(k)(1) Among them, X(k|k-1) is the result of using the state prediction of the previous frame, X(k-1|k-1) is the result of the optimal state of the previous frame, and U(k) is the control of the current frame state amount, if there is no control amount, it is 0; (6-3) Use formula (2) to update the covariance of X(k|k-1), P(k|k-1)＝A·P(k-1|k-1)· ^AT +Q(2), P(k|k-1) is the association corresponding to X(k|k-1) Variance, P(k-1|k-1) is the covariance corresponding to X(k-1|k-1), ^AT represents the transpose matrix of A, and Q is the covariance of the system process; (6-4) Collect the measured value of the current frame state, and combine the predicted value and measured value of the frame state to obtain the optimal estimated value X(k|k) of the current frame state: X(k|k)＝X(k|k-1)+Kg(k)·(Z(k)-H·X(k|k-1))(3) Among them, Kg is Kalman Gain (KalmanGain); Kg(k)=P(k|k-1) ^HT /(H·P(k|k-1)· ^HT +R)(4); (6-5) Constantly update the covariance of K state X(k|k): P(k|k)=(I-Kg(k)·H)P(k|k-1)(5) Where I is a matrix of 1, for a single model and a single measurement, I=1; when the system enters the k+1 state, P(k|k) is P(k|k-1) of the formula (2). 3. A system that adopts the transmission line object detection method as claimed in claim 1, is characterized in that it includes a video camera, the video camera is connected with the video server by RS485, and the video output by the video server is sent into the image acquisition board, and the image acquisition board is connected with the video server The upper computer is connected, the upper computer is connected with the base station, the upper computer is connected with the alarm, and the image acquisition board is also connected with the DSP high-speed compression circuit and the memory.