CN104574407A - Insulator ice coating monitoring method based on dark channel precedence and triangle method - Google Patents

Abstract

The invention discloses an insulator ice coating monitoring method based on dark channel precedence and the triangle method. The insulator ice coating monitoring method mainly includes the following steps that S1 image acquisition is conducted, wherein cameras are erected according to the figure 1 and images of a monitored object is taken; S2 according to dark channel precedence and the triangle method, the distances between the cameras and the monitored object are worked out; S3 according to the camera imaging principle, the actual sizes of all parts of the monitored object in the images are worked out. The power monitoring method is simple, visual, economical and effective.

Description

The covering ice for insulator monitoring method of triangulation method is preferentially conciliate based on dark

Technical field

The invention belongs to the on-line monitoring field of image procossing and insulator, particularly relate to the method being calculated covering ice for insulator situation by picture.

Background technology

China is one of multiple country of ice damage.Transmission line of electricity and insulator affect by ice damage and may occur caving in, and the great electric power accidents such as broken string and insulator arc-over, cause large-area power-cuts.The safe operation of many ice damages regional power system in this serious threat, and therefore suitable monitoring and warning technology is very necessary.The monitoring of current covering ice for insulator mainly contains artificial and sensor two kinds of monitoring modes, but the cost of personal monitoring is higher, and affects comparatively large by subjective factor, the method for Sensor monitoring have again install inconvenient, with the problem such as icing relation complexity.

Can remote monitoring be realized by the method for image procossing, and compare traditional method based on mechanics principle by sensor, more intuitively, economical, effectively.Since nearly ten years, the stakeholder of power domain attempts utilizing the method for image procossing to carry out icing monitoring one after another.But because the profile of insulator affects very large by icing, most trial is all monitor about the icing of transmission line of electricity, and its effect depends on the accuracy of rim detection and Iamge Segmentation to a great extent.

The invention provides a kind of covering ice for insulator situation on-line monitoring method of preferentially conciliating triangulation method based on dark, do not need when light moderate icing to know that the size of insulator just can calculate the situation of its icing, its icing situation can be calculated in conjunction with the dimension information of insulator when severe icing.Because the present invention is without the need to doing rim detection, so more effectively simple.

Summary of the invention

The invention provides a kind of covering ice for insulator situation on-line monitoring method of preferentially conciliating triangulation method based on dark, covering ice for insulator monitoring method provided by the invention mainly comprises the steps:

Preferentially conciliate the covering ice for insulator monitoring method of triangulation method based on dark, comprise the steps:

S1, Image Acquisition, specifically comprise:

S1.1 sets up two video cameras: one is monitoring video camera, and one is auxiliary video camera, and two video cameras and monitoring target regard three points as, and its plane orthogonal formed is in ground;

S1.2 obtains image: take same monitoring target with two video cameras set up in step S1.1, after having obtained image, auxiliary video camera is removable;

S2 calculates the distance x of video camera to monitoring target ₁, specifically comprise:

S2.1 sets up image attenuation model i ^c(x, y)=j ^c(x, y) t (x, y)+A (1-t (x, y));

S2.2 is divided into multiple foursquare fritter image;

S2.3 estimates overall atmosphere light A: get dark i in all described fritters ^darkthe value of (x, y) is positioned at the fritter of front 10%, then selects i wherein ^darkthe fritter that (x, y) intensity is the strongest, its i ^dark(x, y) is A;

S2.4 calculates transmissivity $\tilde{t}(x,y)=1-\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right);$

S2.5 calculates the distance ratio of two video cameras to detected object $k=\frac{x_2}{x_1}=\frac{1}{n}\underset{s=1}{\overset{n}{\mathrm{\Σ}}}\frac{\ln \left(t_2(x_s,y_s)\right)}{\ln \left(t_1({x_s}^{\′},{y_s}^{\′})\right)};$

S2.6 calculates the monitoring distance x of video camera to monitoring target ₁, $x_1=\frac{d^2}{1+k^2-2k\cos ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)};$

S3 monitors the icing situation of insulator, specifically comprises:

S3.1 calculates the relation of camera imaging size and pixel: picture element density

S3.2 calculates the physical size of the monitoring target various piece (as icing) in picture: according to and can in the hope of picture element density by S3.1, again because therefore physical size can be tried to achieve;

S3.3 carries out angle modification: if the result in step S3.2 is that reference will divided by cos θ with longitudinal direction ₁revise, θ ₁it is the angle of monitoring video camera and monitoring target.

Further, the camera arrangements triangular established law of described step S1.1, is specially: two video cameras and monitoring target regard three points as, and its plane orthogonal formed is in ground, and monitoring video camera is respectively θ with auxiliary with video camera and ground angle ₁and θ ₂, formed by monitoring target and two video camera lines, angle is θ ₃=θ ₂-θ ₁, the distance between two video cameras is d, and its distance to monitoring target is respectively x ₁, x ₂.

Further, calculate the method for overall atmosphere light A in described step S2.3, be specially: because occurring in nature exists shadow phenomenon, for the original image j not having to decay ^ceach fritter in (x, y), at least has some pixel values close to 0 at a Color Channel, namely $j^{\mathrm{dark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b\}}{\min }{j}^c(x,y)\right)\→0,$ So by obtaining the rear image i of decay ^cthe dark i of each fritter of (x, y) ^dark(x, y), its value is approximately atmosphere light A, for increasing accuracy, gets i in all fritters ^dark(x, y) value is positioned at the fritter of front 10%, then selects i wherein ^darkthe fritter that (x, y) intensity is the strongest, its i ^dark(x, y) is A.

Further, calculate the method for transmissivity in described step S2.4, be specially: obtained by the image attenuation model in S2.1 $\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right)=\tilde{t}(x,y)\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{j^c(x,y)}{A}\right)+1-\tilde{t}(x,y),$ Obtained by step S2.3 $j^{\mathrm{dark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b\}}{\min }{j}^c(x,y)\right)\→0,$ $\tilde{t}(x,y)=1-\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right).$

Further, ask two video cameras to the method for monitoring target distance ratio in described step S2.5, be specially: in homogeneous atmospheric environment, transmissivity t (x, y)=e ^{-β d (x, y)}, wherein d (x, y) represents the scene depth of point (x, y), so wherein t ₁(x _s', y _s') and t ₂(x _s, y _s) be respectively plate center (x captured by two cameras _s', y _s') and point (x _s, y _s) transmissivity, represent in actual monitoring object not to be white point.

Further, ask monitoring video camera to the method for monitoring target distance in described step S2.6, be specially: obtain d by the cosine law ²=x ₁ ²+ x ₂ ²-2x ₁x ₂cos θ ₃, again because being obtained by step S3 $k=\frac{x_2}{x_1},$ So $x_1=\frac{d^2}{1+k^2-2k\cos ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)}.$

Further, monitoring method described in described S3, is specially:

S3.1 calculates the relation of camera imaging size and pixel: the given parameters according to camera obtains picture element density, if it is r, so

S3.2 calculates the physical size of the monitoring target various piece (as icing) in picture: the imaging relations according to camera obtains, again because imaging focal length can be shown in the parameter of captured image, camera can be obtained by S2.6 to monitoring target distance, so the physical size of monitoring target various piece can be obtained by picture, for slight and moderate icing, directly can calculate, for the situation that icing is serious, the thickness that insulator adds two-layer icing can be calculated, obtaining the physical size of insulator by tabling look-up, subtracting each other the thickness being icing again divided by 2;

S3.3 carries out angle modification: be face upward bat owing to being video camera, if the result in step S3.2 is with reference to will divided by cos θ with longitudinal direction ₁revise, θ ₁it is exactly the angle of monitoring video camera and monitoring target.

Further, said method also comprises covering ice for insulator early warning, when the icing block obtained is greater than insulator ability to cope with the exigency, just sends warning message.

Compared with prior art, beneficial effect of the present invention comprises:

The present invention can realize the object of early warning by the real conditions of reduction monitoring target various piece (as icing), can also be used for the vertical arc etc. of monitoring overhead cable except insulator, not only safety but also economical and effective.

Accompanying drawing explanation

Fig. 1 is the position view of two video cameras and monitoring target.

Fig. 2 is a kind of workflow diagram of preferentially conciliating the covering ice for insulator monitoring method of triangulation method based on dark of the present invention.

Embodiment

Below in conjunction with accompanying drawing, enforcement of the present invention is described further, but embodiments of the present invention are not limited thereto, if it is noted that there be not special detailed description in detail to point out below, be all that those skilled in the art can refer to existing techniques in realizing.

Preferentially conciliate a covering ice for insulator monitoring method for triangulation method based on dark, comprise the steps:

S1 Image Acquisition, specific as follows:

S1.1 sets up two video cameras: one is monitoring video camera, one is auxiliary video camera, and removable in observation process, two video cameras and monitoring target regard three points as, its plane orthogonal formed is in ground, and monitoring video camera is respectively θ with auxiliary with video camera and ground angle ₁and θ ₂, formed by monitoring target and two video camera lines, angle is θ ₃=θ ₂-θ ₁, the distance between two video cameras is d, and its distance to monitoring target is respectively x ₁, x ₂, as shown in Figure 1;

For ensure two video cameras and monitoring target same plane can with light beam by video camera and monitoring target line to determine position, measurement of angle can by angulometer;

S1.2 obtains image: take same monitoring target with two video cameras set up in S1.1, after having obtained image, auxiliary video camera is removable;

S2 calculates the distance x of video camera to monitoring target ₁, specific as follows:

S2.1 sets up image attenuation model i ^c(x, y)=j ^c(x, y) t (x, y)+A (1-t (x, y)): wherein i ^c(x, y) is the pixel value of the upper point (x, y) of passage c (c ∈ { r, g, b}) in the image (image due to the particulate decay in air) observed, j ^c(x, y) passage c (c ∈ { r in the original image not having to decay is represented, g, b}) pixel value of upper point (x, y), t (x, y) point (x is represented, y) transmissivity, does not namely arrive the number percent of the light of camera through scattering, A represents overall atmosphere light;

S2.2 is divided into many foursquare fritters (concrete value is according to the pixel size of picture) image, and the fritter of such as 5 × 5, is designated as Ω (x, y), approximately thinks that t (x, y) in each fritter is for invariable, is designated as

S2.3 estimates overall atmosphere light A: because occurring in nature exists shadow phenomenon, for the original image j not having to decay ^ceach fritter in (x, y), at least has some pixel values close to 0 at a Color Channel, namely $j^{\mathrm{dark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b\}}{\min }{j}^c(x,y)\right)\→0,$ As long as so image i after obtaining decay ^cthe dark i of each fritter of (x, y) ^dark(x, y), its value can be similar to regards overall atmosphere light A as, for increasing accuracy, gets i in all fritters ^dark(x, y) value is positioned at the fritter of front 10%, then selects i wherein ^darkthe fritter that (x, y) intensity is the strongest, its i ^dark(x, y) is A;

S2.4 calculates each square tiles can be obtained by the image attenuation model in S2.1 $\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right)=\tilde{t}(x,y)\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{j^c(x,y)}{A}\right)+1-\tilde{t}(x,y),$ Obtained by S2.3 $j^{\mathrm{dark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b\}}{\min }{j}^c(x,y)\right)\→0,$ So $\tilde{t}(x,y)=1-\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right);$

S2.5 calculates the distance ratio of two video cameras to detected object in all consistent atmospheric environment, transmissivity t (x, y)=e ^{-β d (x, y)}, wherein d (x, y) represents the scene depth of point (x, y), so wherein t ₁(x _s', y _s') and t ₂(x _s, y _s) be respectively plate center (x captured by two cameras _s', y _s') and point (x _s, y _s) transmissivity, represent in actual monitoring object not to be white point;

S2.6 calculates the monitoring distance x of video camera to monitoring target ₁: as shown in Figure 1, obtain d by the cosine law ²=x ₁ ²+ x ₂ ²-2x ₁x ₂cos θ ₃, again because being obtained by S3 so $x_1=\frac{d^2}{1+k^2-2k\cos ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)};$

S3 monitors the icing situation of insulator, specific as follows:

S3.1 calculates the relation of camera imaging size and pixel: the given parameters according to camera obtains picture element density, if it is r, so

S3.2 calculates the physical size of the monitoring target various piece (as icing) in picture: the imaging relations according to camera obtains, can in the hope of picture element density by S3.2, and again because imaging focal length can be shown in the parameter of captured image, camera can be obtained by S2.6 to monitoring target distance, so the physical size of monitoring target various piece can be obtained by picture, as for slight and moderate icing, directly can calculate, for the situation that icing is serious, the thickness that insulator adds two-layer icing can be calculated, obtaining the physical size of insulator by tabling look-up, subtracting each other the thickness being icing again divided by 2;

S3.3 carries out angle modification: be face upward bat owing to being video camera, if the result in S3.2 is with reference to will divided by cos θ with longitudinal direction ₁revise, θ ₁it is exactly the angle of monitoring video camera and monitoring target.

Above-described embodiment can as better embodiment of the present invention; but embodiments of the present invention are not limited by the examples; other any do not deviate from make under Spirit Essence of the present invention and principle change, modification, substitute, combine, simplify; all should be the substitute mode of equivalence, be included within protection scope of the present invention.

Claims (8)

1. preferentially conciliate the covering ice for insulator monitoring method of triangulation method based on dark, it is characterized in that comprising the steps:

S1, Image Acquisition, specifically comprise:

S1.1 sets up two video cameras: one is monitoring video camera, and one is auxiliary video camera, and two video cameras and monitoring target regard three points as, and its plane orthogonal formed is in ground;

S1.2 obtains image: take same monitoring target with two video cameras set up in step S1.1, after having obtained image, auxiliary video camera is removable;

S2 calculates the distance x of video camera to monitoring target ₁, specifically comprise:

S2.1 sets up image attenuation model i ^c(x, y)=j ^c(x, y) t (x, y)+A (1-t (x, y));

S2.2 is divided into multiple foursquare fritter image;

S2.3 estimates overall atmosphere light A: get dark i in all described fritters ^darkthe value of (x, y) is positioned at the fritter of front 10%, then selects i wherein ^darkthe fritter that (x, y) intensity is the strongest, its i ^dark(x, y) is A;

S2.4 calculates transmissivity $\tilde{t}(x,y)=1-\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right);$

S2.5 calculates the distance ratio of two video cameras to detected object $k=\frac{x_2}{x_1}=\frac{1}{n}\underset{s=1}{\overset{n}{\mathrm{\Σ}}}\frac{\ln \left(t_2(x_s,y_s)\right)}{\ln \left(t_1({x_s}^{\′},{y_s}^{\′})\right)};$

S2.6 calculates the monitoring distance x of video camera to monitoring target ₁, $x_1=\frac{d^2}{1+k^2-2k\cos ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)};$

S3 monitors the icing situation of insulator, specifically comprises:

S3.1 calculates the relation of camera imaging size and pixel: picture element density

S3.2 calculates the physical size of the monitoring target various piece in picture: according to and can in the hope of picture element density by S3.1, again because of therefore physical size can be tried to achieve;

S3.3 carries out angle modification: if the result in step S3.2 is that reference will divided by cos θ with longitudinal direction ₁revise, θ ₁it is the angle of monitoring video camera and monitoring target.

2. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 1, it is characterized in that, the camera arrangements triangular established law of described step S1.1, be specially: two video cameras and monitoring target regard three points as, its plane orthogonal formed is in ground, and monitoring video camera is respectively θ with auxiliary with video camera and ground angle ₁and θ ₂, formed by monitoring target and two video camera lines, angle is θ ₃=θ ₂-θ ₁, the distance between two video cameras is d, and its distance to monitoring target is respectively x ₁, x ₂.

3. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 1, it is characterized in that, calculate the method for overall atmosphere light A in described step S2.3, be specially: because occurring in nature exists shadow phenomenon, for the original image j not having to decay ^ceach fritter in (x, y), at least has some pixel values close to 0 at a Color Channel, namely $j^{\mathrm{bark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b\}}{\min }{j}^c(x,y)\right)\→0,$ So by obtaining the rear image i of decay ^cthe dark i of each fritter of (x, y) ^dark(x, y), its value is approximately atmosphere light A, for increasing accuracy, gets i in all fritters ^dark(x, y) value is positioned at the fritter of front 10%, then selects i wherein ^darkthe fritter that (x, y) intensity is the strongest, its i ^dark(x, y) is A.

4. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 1, is characterized in that, calculate the method for transmissivity, be specially in described step S2.4: obtained by the image attenuation model in S2.1 $\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right)=\tilde{t}(x,y)\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{j^c(x,y)}{A}\right)+1-\tilde{t}(x,y),$ Obtained by step S2.3 $j^{\mathrm{dark}}(x,y)=\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c\∈\{r,g,b}{\min }{j}^c(x,y)\right)\→0,$ $\tilde{t}(x,y)=1-\underset{(x_0,y_0)\∈\mathrm{\Ω}(x,y)}{\min }\left(\underset{c}{\min }\frac{i^c(x,y)}{A}\right).$

5. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 1, it is characterized in that, ask two video cameras to the method for monitoring target distance ratio in described step S2.5, be specially: in homogeneous atmospheric environment, transmissivity t (x, y)=e ^{-β d (x, y)}, wherein d (x, y) represents the scene depth of point (x, y), so wherein t ₁(x _s', y _s') and t ₂(x _s, y _s) be respectively plate center (x captured by two cameras _s', y _s') and point (x _s, y _s) transmissivity, represent in actual monitoring object not to be white point.

6. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 2, is characterized in that, asks monitoring video camera to the method for monitoring target distance, be specially: obtain d by the cosine law in described step S2.6 ²=x ₁ ²+ x ₂ ²-2x ₁x ₂cos θ ₃, again because being obtained by step S3 so $x_1=\frac{d^2}{1+k^2-2k\cos ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)}.$

7. covering ice for insulator monitoring method of preferentially conciliating triangulation method based on dark according to claim 1, it is characterized in that, monitoring method described in described S3, is specially:

S3.1 calculates the relation of camera imaging size and pixel: the given parameters according to camera obtains picture element density, if it is r, so

S3.2 calculates the physical size of the monitoring target various piece in picture: the imaging relations according to camera obtains, again because imaging focal length can be shown in the parameter of captured image, camera can be obtained by S2.6 to monitoring target distance, so the physical size of monitoring target various piece can be obtained by picture, for slight and moderate icing, directly can calculate, for the situation that icing is serious, the thickness that insulator adds two-layer icing can be calculated, obtaining the physical size of insulator by tabling look-up, subtracting each other the thickness being icing again divided by 2;

S3.3 carries out angle modification: be face upward bat owing to being video camera, if the result in step S3.2 is with reference to will divided by cos θ with longitudinal direction ₁revise, θ ₁it is exactly the angle of monitoring video camera and monitoring target.

8. according to claim 1 ?method described in 7 any one, it is characterized in that, also comprise covering ice for insulator early warning, when the icing block obtained is greater than insulator ability to cope with the exigency, just send warning message.