1. TECHNICAL FIELD OF INVENTION

The present invention relates generally to a method and system for vegetation encroachment monitoring comprises of at least one image capturing device which is integrated with wireless sensor network (WSN) to form a wireless multimedia sensor network (WMSN) to transmit the acquired image data to monitoring station for image processing. A triangulation based method is used to determine the height of vegetation as well as the distance between said vegetation and the object of interest for vegetation encroachment monitoring.

2. BACKGROUND OF THE INVENTION

Vegetation encroachment monitoring is particularly important as a preventive measure especially in electric utility system in order to prevent service interruptions as well as safety hazards which associated with trees. Oftentimes, the bad weather conditions such as rain, wind and storm would cause the falling trees or branches strike to the power line sag and hence causing power lines out of service. Therefore, vegetation encroachment monitoring is crucial to ensure the safety of overhead lines at anytime against endangering encroachments.

Traditionally, the vegetation monitoring have to rely on regular visual field survey, whereby a team is deployed to inspect the power lines either by pole climbing or using vehicles. For pole-climbing, linesmen climb up a pole with computer having information regarding the components at that pole. This method is less accurate due to judgmental errors by humans for vegetations that appear to be in safe clearances but could be dangerous during bad weather conditions.

Other methods of vegetation monitoring includes aerial inspection such as helicopter surveillance, videography and air-borne LiDAR scanners. Helicopters mounted with surveillance cameras are used to inspect the region with transmission line network. This method is ambiguous for a non- uniform terrain due to changing perspective of target via random motion of camera in vertical direction. The airborne LiDAR scanning is an optical remote sensing technique that uses the scattering of light to identify the range and other information of the distant object. However, uneven hovering of air-borne vehicle causes ambiguities in the data recorded by LiDAR and consequently the software used for 3D tracking of transmission line can produce ambiguous model of the scene. Furthermore, the usage of LiDAR is very expensive though a more accurate result may be obtained. Whereas aerial videography technology captures images of ground from an elevated position. It is extensively used for the agricultural problems and so for vegetation monitoring. However, low flight altitude (<500m), difficulty in obtaining flight permission, high cost and low accuracy are the main disadvantages if aerial videography is used.

It would hence be extremely advantageous if the above shortcoming is alleviated by having a method and system for vegetation encroachment monitoring that comprises of at least one image capturing device integrated with wireless sensor network (WSN) to form WMSN to transmit the acquired image data to monitoring station for image processing. The methodology of the present invention is a non-airborne method that uses image processing platform to monitor the vegetation encroachment within and outside right-of-ways.

3. SUMMARY OF THE INVENTION

Accordingly, it is the primary aim of the present invention to provide a method and system for vegetation encroachment monitoring whereby the height of encroached vegetation with respect to object of interest can be identified to determine whether the vegetation has entered the danger zone. It is yet another objective of the present invention to provide a method and system for vegetation encroachment monitoring whereby the distance between the object of interest and the vegetation can be identified to ensure a safe distance between the object of interest and the vegetation.

It is yet another objective of the present invention to provide a method and system for vegetation encroachment monitoring to identify dangerous vegetation which is outside right-of-ways that appear to be safe during normal atmospheric conditions but could be dangerous during bad weather conditions.

It is yet another objective of the present invention to provide a method and system for vegetation encroachment monitoring which is cost-effective, lesser time consuming and more accurate results obtained.

Other and further objects of the invention will become apparent with an understanding of the following detailed description of the invention or upon employment of the invention in practice.

According to a preferred embodiment of the present invention there is provided, A method for vegetation encroachment monitoring relative to an object of interest comprising:- i. acquiring at least one reference image and at least one scene image of the object of interest; ii. pre-processing of said reference image and said scene image comprising of the sub-steps of filtering and identifying its weather conditions; iii. identifying horizontal lines and vertical lines of said object of interest to determine horizontal and vertical thresholds for vegetation encroachment monitoring; iv. detecting the encroached vegetations in the scene image with respect to reference images; v. estimating width and height of encroached vegetation, and distance between said object of interest and the encroached vegetation using depth from triangulation methodology for both within and outside the right-of-ways to determine the status of said encroached vegetation. characterized in that said step of detecting the encroached vegetations is done by sub-steps of: a. executing object tracking by subtracting said reference images from said scene images; b. using binarization for segmentation of the resulting data; c. using said horizontal and vertical thresholds to discard unwanted illumination effects within and outside right-of-ways of said object of interest; d. executing morphological operators to remove uneven appearances.

In a second embodiment of the present invention there is provided,

A system for vegetation encroachment monitoring relative to an object of interest comprising: i. at least one image capturing device to acquire at least one reference image and at least one scene image of the object of interest; ii. at least one electronic processing means for processing and analysing said acquired image; characterized in that said image capturing device is integrated with wireless sensor network (WSN) to form a wireless multimedia sensor network (WMSN) to transmit the acquired image data to monitoring station for image processing by said electronic processing means; said image capturing device is mounted anywhere as long as the image for said object of interest is able to be acquired.

4. BRIEF DESCRIPTION OF THE DRAWINGS

Other aspects of the present invention and their advantages will be discerned after studying the Detailed Description in conjunction with the accompanying drawings in which:

FIG. 1 shows a flow diagram outlining the general steps for the vegetation encroachment monitoring.

FIG. 2-A shows an exemplary of implementing initialization process for reference image and scene image acquisition. FIG. 2-B shows an exemplary of implementing filtering and weather condition identification process.

FIG. 3-A and FIG. 3-B show the identification of horizontal lines and vertical lines of transmission pole.

FIG. 4-A to FIG. 4-C show the coordinate arrangement of 3x3 masking coefficients with horizontal mask (FIG. 4-B) and vertical mask (FIG. 4-C) to filter the reference image.

FIG. 5-A to FIG. 5-C show the placement of horizontal and vertical thresholds on subtracted-binarized scene image to identify zone of encroached vegetation within and outside right-of-ways.

FIG. 6 shows a flow diagram of depth from triangulation algorithm.

FIG. 7-A and FIG. 7-B show a distance between the far away pole and the vegetation which is located at the back side or in front of the far away pole.

FIG. 7-C shows a model for tree on a non-flat surface.

FIG. 7-D shows a model for computing height of the tree. FIG. 8 shows a model to compute the distance between the Object of Interest (for example transmission line) and the encroached vegetation (outside the Object of Interest).

FIG. 9 shows the integration of WMSNs communicating in multi-hop manner after monitoring at a particular node.

5. DETAILED DESCRIPTION OF THE DRAWINGS

In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those or ordinary skill in the art that the invention may be practised without these specific details. In other instances, well known methods, procedures and/or components have not been described in detail so as not to obscure the invention.

The invention will be more clearly understood from the following description of the methods thereof, given by way of example only with reference to the accompanying drawings, which are not drawn to scale.

Referring now to FIG. 1, there is shown a flow diagram outlining the general steps for the vegetation encroachment monitoring relative to an object of interest whereby at least one image capturing device is integrated with wireless sensor network (WSN) on the object of interest to form a wireless multimedia sensor network (WMSN) for data acquisition as well as transmitting the acquired data for data processing. Alternatively, said acquired image data can be processed onsite before being transmitted by said WMSN. Said object of interest includes transmission line (including transmission Pole), power line pole, distribution line pole, communication cable pole, and construction structures. Hereinafter, transmission line pole will be used as an example of object of interest to describe on the method for vegetation encroachment monitoring. Said methodology of the present invention also can be used for identification of nearby vegetation such as trees, shrubs, plants, etc. which is outside right-of-ways within the specified zone that appear to be safe during normal atmospheric conditions but could be dangerous during bad weather conditions.

From the flow diagram, the first step for the methodology of tlie present invention is initialization by acquiring reference image and scene image of tlie object of interest (110). Said reference image are acquired when vegetation is trimmed to its predetermined minimum level which is a zone out of danger within and outside transmission lines right-of-ways.

FIG. 2-A shows an exemplary of implementing said initialization process for reference image and scene image acquisition. Said reference images are acquired every five minutes (from 9am to 5pm) for initial ten days after vegetation is trimmed. Typically there will be no changes in the scene as vegetation growth is relatively a slow process regardless of the weather conditions. The criteria for capturing said reference image varies with weather conditions. For example, Malaysia which is a tropical country and weather changes abruptly but temperature remains warm throughout the year. This is unlike other parts of the world where the weather is hot in . summer and cold in winter. The images acquired after ten days are considered as scene images that will be analyzed by at least one electronic processing means for detection of vegetation encroachments and determining the status of vegetation encroachments.

In the second step, pre-processing is done which includes filtering and weather condition identification (120) in tliose image frames. FIG. 2-B shows an exemplary of implementing said filtering and weather condition identification process. Reference image and scene image of the object of interest are filtered by a median filter of window size 3x3 or any other filtering to remove unwanted noise. Gamma correction is used to adjust the brightness in those images. The weather conditions such as sunny, rainy and foggy are identified in all the reference and scene images in order to only consider the image frames with appropriate illumination (210). Pattern recognition techniques such as Principal Component Analysis and "Dark Channel Prior Model" are used to extract features for the automated classification of weather (illumination) conditions in the images. The rainy images are identified by detecting the rain drops on windshield of the camera mounted on transmission pole by extracting the raindrop features such as uniform in shape and refraction of light. Principal Component Analysis (PCA) is used to recognize the pattern of raindrops on the windshield of camera by matching raindrop frames with the incoming reference or scene frames. At first M frames with raindrops on the windshield, each of Width (W) and Height (H) are captured and represented by one-dimensional vectors, that are normalized to unit vectors given as: x; = (xj, xi, ..., XN) ^T , where N = W x H. Let X=[ xi, x ₂ ,..., XM] be a matrix of M randomly selected vectors from the test images, and its covariance matrix be Z=XX ^T . The largest eigenvalues Q of Z and its corresponding eigenvectors {ei,e ₂ ,...,eQ_} are computed. The rain detection for any normalized reference or scene image is done by computing its degree of similarly with the eigenvectors (raindrops in template frames). If K is the test image, then S(K) is given by:

S K ) - ∑(A\ c )

(1)

The frame is detected as rainy if S(K) is greater than a threshold by computing the similarity between the eigenvectors. The foggy images are identified by using the "Dark Channel Prior Model" to measure the fog level in an image. This dark channel is computed for first prior reference frame which will be a clear image observed visually at base- station. The image of dimension W*-H is further divided into K*L patches. The darker intensity value in each patch is obtained by finding the minimum of all. After that the dark channel prior of this clear image is given by:

Where n is the number of patches. After that, the dark channel of next new incoming reference or scene image is computed as:

The differences between the two dark channels of the corresponding patches is calculated. The average difference is computed as:

(4)

The frame is detected as foggy if D is greater than a threshold. Then, the reference image that is free from fog and rain and having sufficient illumination is sorted as a dataset at monitoring station such as base station with respect to date, time, weather and contrast. The scene images acquired are also filtered and detected to be either rainy, foggy or low illumination. Scene image is then selected having the average intensity value ranging between the values of Highest Contrast (HC) and Lowest Contrast (LC) from a dataset of reference images (220). After the identification of acceptable scene image, at least two to three reference images are extracted from the entire dataset having the same illumination (weather) as captured scene image. This is done by using PCA which retains the variation in data as much as possible (230). Each reference/ scene image taken is an 8-bit array of size W*H, which can also be considered as a column vector having dimension (W*H)xl. Let h, h, h,..., be the training dataset of reference images acquired for ten days in an outdoor environment under varying illuminations from 9am in morning up to 5pm in evening. Next compute the mean of the frames in the training dataset, and difference of each image frame (of the dataset) from the mean vector to form a matrix A = [Φι, Φ?, Φ3 _/ ..., Φζ], where (W*H)xZ. After that find the covariance matrix C = AA ^T , and compute its the eigen vectors (vi, in..., vx) and eigen values (A?, A ₂ ,..., Ax). The dimensionality is reduced from X to Y space by keeping only the eigen vectors corresponding the largest eigen values:

Represented in form of a vector containing the weights as:

Equation (6) describes the contribution of each reference frame in dataset, and is used to identify the illumination pattern of an input scene image with encroached vegetations. Repeat this procedure to find the vector containing the weights of scene image as:

The reference image which is similar to scene image is identified computing which is the Euclidean distance between the weights of scene and reference images of training dataset. Two to three reference images that are closer or having same illumination to scene images with minimum distances are thus can be sorted out (240).

The third step in the vegetation encroachment monitoring is the identification of transmission pole (130). Horizontal and vertical lines of far away transmission pole is identified by implementing Hough Transform to said reference image as shown in FIG. 3-A and FIG. 3-B. to determine the horizontal and vertical thresholds for vegetation encroachment monitoring. Hough Transform is used for feature detection especially straight lines in image processing and it states that n number of co-linear points on a straight line in x-y plane correspond to n number of straight lines passing through a single point in parameter space. For computation reasons, Rough (p) - theta (Θ) space is used in Hough Transform instead of Parameter space. The detection of horizontal lines and threshold is given below as: i. Filter the reference image (x,y,ti) with horizontal mask as in FIG. 4-B to find out the candidates for horizontal lines, where is any time instant.

I r

R(x, y) = w(x,y)® I _R (x,y) =∑∑ w„ _. jA* + + ^m ) (8)

M = - lm = -l

Where R(x,y) is the resulting filtered image, and xv(x,y) shows the masking coordinates as in FIG. 4-A used to filter the reference image (x,y,ti). ii. After that detect the edges by using canny edge detector.

BW — Apply Canny Edge detector on R(x,y) Where BW is the output binary image. iii. Apply Hough Transform to BW to identify horizontal lines in the image.

L*— Apply hough transform on BW iv. Only detect horizontal lines of far away pole and a suitable horizontal threshold to monitor the vegetations, the pseudo code is expressed below: Initialize Program

PROCEDURE GET (L, h(x,y))

1. FOR g=l to length (L)

2. f{g) <— An array of row indexes for each horizontal line

3. END

4. [Sx, S _y ] i— Dimensions oflR(x,y).

5. horz center <— S _¾ 2.

6. upper _range <— horz_center - Sx 4.

7. lower _range <— liorz_center + Sx/4.

8. [a,b] <— FIND (upper _range < f && lower _range >f)

9. IF (b not equal 1) THEN

10. hjines «-/(& _;

11. hjresh MAXIMUM (hjines).

12. RETURN hjresh

13. ELSE "Display no Line detected".

14. End of Program

h_tresh is the horizontal threshold which shows the detected horizontal line of the pole as in FIG. 3-A.

Similarly, the vertical lines and two vertical thresholds of the same pole is identified by taking the Hough Transform of the reference image to monitor the encroachments witliin and outside right-of-ways. Vertical mask as shown in FIG. 4-C is used to determine candidates for vertical lines. tresh <— Two -vertical thresholds [vi where vjtresh consists of two vertical thresholds vi and VR respectively, which represent the left and right most vertical lines (threshold) of the pole as shown in FIG. 3-B. The fourth step in the vegetation encroachment monitoring is to detect the change in the height of the encroached vegetations (140) in the scene image Is(x,y,tj) with respect to reference images (x,y,ti) by using object tracking method. Where t _j is any time instant depending upon the specie of encroaching vegetations. Object tracking is done by filtering the scene image with Laplacian kernel to enhance the edges, and subtracting the reference image from the scene image by using motion based segmentation that uses binarization as given below:

Where T is the threshold for binarization and its value is set to average intensity value of sorted reference images. The subtracted-binarized image would include vegetations that are encroached in the particular duration after the reference image has been taken. Typically effects of uneven illumination or noise which appeared in the binarized image due to the random changes of lightening in the scene would create a problem in obtaining an accurate result. To overcome this, horizontal and vertical thresholds are used to discard illumination effects outside right-of-ways. Closing (dilation followed by erosion) of an image by 'square' element of size 2 is used as a morphological operator within right-of-ways to remove uneven appearances. The level of excess vegetations is determined by comparing the pixel values of binarized reference image with the subtracted-binarized image. This is done by placing few horizontal monitoring thresholds within and outside right-of-ways on the after-subtraction binarized image given below in the pseudo code:

Initialize Program

PROCEDURE GET (hs(x,y), hjresh)

1. [XRS, YRS] <— Dimension (IRS(X,JI )).

2. horz_pixels <— XRS - h_tresh.

3. Division <— horz_pixels / 8.

4. I +- 1.

5. tres (I) — hjresh.

6. WHILE (I<=7) DO

7. tres (1+1) *— tres (I) + Division.

8. I *- J+L

9. END 10. End of Program

The binarized image hs(x,y) is further divided into eight monitoring thresholds as in the above pseudo code. FIG. 5-A and FIG. 5-B shows eight thresholds on the subtracted-binarized and scene image. Thus, by using those monitoring zones, the level of overgrown vegetations within and outside right-of-ways can be identified.

In real-time environment trees or shrubs may appear to be in safe zone with respect to transmission lines outside right-of-ways, but may interfere or fall into danger zone during bad weather. Therefore, vegetation outside right-of- ways are also monitored by tracking the level of vegetations to identify the zone and also determine whether the vegetation is at the safe distance from the transmission lines. The distance identification is carried out by placing the vertical monitoring thresholds on left and right side right-of-ways of transmission line and after that comparing the level of vegetation with the distance of vegetation from the transmission lines. For example, if both the level of vegetation and distance from transmission line is also in danger zone then the overall status of the vegetation would be dangerous and vice-versa as in FIG. 5-C.

In the current step, said level of encroached vegetation is identified within and outside the Object of Interest. A decision is made if the encroached vegetation has exceeded above the maximum allowable level then the algorithm will proceed further to the next step else if the encroached vegetation is within the safe zone then it will return back to the first step waiting for the scene (x,y,tj) image frame after a certain period of time. The fifth step in the vegetation encroachment monitoring is estimating width and height of encroached vegetation, and distance between said object of interest and the encroached vegetation using depth from triangulation metliodology for both within and outside the right-of-ways (150). Depth from Triangulation (DfT) is used to determine the distance between the transmission pole where camera is mounted and trees (within and outside right-of-ways). This is carried out by firstly extracting the overgrown vegetations in the scene by using background-subtraction. Then the algorithm •identifies the ground location of vegetations in the scene because the centroid does not represent the true location of the image in 3D world coordinates. Said algorithm assumes having a camera with Wireless Sensor Network (WSN) or any other transmission means installed at a known height and a known vertical angle on the transmission pole. In this model, the vegetation coordinates (X, Y, Z) are computed using some geometrical information which is the input parameters comprises of camera extrinsic parameters, i.e. camera height (h) and camera vertical angle (Θ), camera intrinsic parameters, i.e. vertical field of view (FO V _V ) and horizontal field of view (FO VH) as well as image parameters, i.e. image size (W, H) and tree location p(i,j). FIG. 6 shows a flow diagram with geometrical information required to compute the vegetation coordinates.

Referring now to FIG. 7- A and FIG. 7-B, there is shown a trigonometry model of a vegetation in the scene at location p(i,j). The image size is with widtli (W) and height (H). The rotation angle (Φ) and vertical angle (Ψ) are computed for the tree located at point p(i,j) as given below:

Then the distances X and Y are computed using these two angles as given below:

X = Y x tan( )

(13) The three dimensional coordinates of the tree at point p(i,j) are then computed assuming that the center of coordinates is beneath the camera.

L ² = X ² + Y ² (14)

Z ² = h ² + L ² = h ² + X ² + Y ² (15)

Z = A7l + tan ² (v/) sec ² (^) (17) where Z is the depth of field or the actual distance between the vegetation and the camera.

In FIG. 7-A and FIG. 7-B, transmission poles are shown whereby the tree can be either on backside of the far away pole or in the front. This can be determined by calculating the distance between the tree and the camera. If the obtained value is positive, the tree is at the backside of the pole. If the obtained value is negative, the tree is in front of the far away pole.

Referring now to FIG. 7-C, there is shown a model for tree on a non-flat surface. A true location of tree is at point (BP) while it is seen in tl e image at point (C) of the ground which is not the ground location of the free. A correct trigonometry relationship cannot be established if the surface is non-flat surface and hence a wrong position is recorded for the tree. If the height of the uneven surface is known or can be estimated, the true depth could be computed by taking the camera height from the tip of the uneven surface. Thus, the present methodology can be implemented on the flat surfaces and non-flat surfaces with known height.

Referring now to FIG. 7-D, there is shown a model to compute the height of the tree. The present methodology can be used to compute the depth and the 3D location of tree that are located on the ground. The height of the tree can be computed by constructing a geometrical model using the bottom point (BP) and the top point (TP) of the object. Then similar triangles in the scene are used to compute the height. The ground distance (Z ) is computed using point (BP) and the ground distance (Z2) is computed using the triangulation algorithm from point (TP). Then, by using similar triangles, the object height (X) is computed using equation (18).

Z2 - Z\

X = h x (18)

Z2

In real-time environment trees or shrubs may appear to be in safe zone with respect to transmission lines outside right-of-ways, but may interfere or fall into danger zone during bad weather. Therefore, vegetation outside right-of- ways (left and right side right-of-ways) are also monitored by tracking the level of vegetations as mentioned in fourth and fifth step to identify the zone (dangerous, high, medium etc.) and also determine whether the vegetation is at the safe distance from the transmission lines. After determining the zone of the encroached vegetation with respect to the transmission lines and poles, a model as shown in FIG. 8 is used to compute the distance between the transmission lines and the encroached vegetation outside right-of-ways.

As shown in FIG. 8, the distance from the encroached vegetation to the transmission lines can be calculated by determining the coordinates of the intersection between the two lines.

Where (XJ,YI) = Coordinates of the transmission pole 1.

(X2, Yi) = Coordinates of the transmission pole 2.

(Χν,Υν) = Coordinates of the vegetation/ tree outside right-of-ways.

(Χρ, Υρ) = Coordinates on the transmission lines.

Dt- v = Distance between the vegetation and pole 1.

D2-V = Distance between the pole 2 and vegetation/ tree.

Dp- v = Distance between the vegetation and transmission lines.

DI-P = Distance between the pole 1 and intersection coordinates on transmission lines.

DTOTAL = Distance between two transmission poles. The equation for transmission line is given as:

"

-^ = m _min X-X _{ ) (20)

Equation for estimating perpendicular distance between the encroached vegetation (outside the transmission lines) and the line connecting the two transmission poles is:

θ _ν =θ _ΤΧΗη ,-90> . »'v =tan(i%,) ₍₂₁₎

Y-Y _v =m _v {X-X _v ) (22)

The intersection of the two lines is:

Y _P -Y _v =m _v {X _P -X _v ) (24)

Subtracting Equation (23) from (24), we get: Y _V -Y = {m _mine - m, ).X _P + m _v X _v - m _TXline X (25) The coordinates (X _P , Y _p ) of the respective point on the transmission lines is :

_Xp = { ^γ ν - ^γ \)-( ^}η ν ^χ ν -™ _τ χ,^ ^χ \) ^26)

Thus, the distance between outside vegetation and the overhead lines DP-V is

The distance between the outside vegetations and the far away pole is:

Where D _P _ ₂

The coordinates (Χι,Υι), { irYi) and(Xv, ν,Όι-ν) can be found by using above Triangulation method. By using these coordinates and the respective slopes mv and mrxune, the coordinates (Χρ,Υρ) can be calculated. The overall status of the vegetation is found by comparing the level (height) of vegetation, and the distance from transmission lines. For example, if the level of encroached vegetation outside the transmission lines is in danger zone and distance of vegetation from the transmission lines is also in danger zone then the overall status of the vegetation encroachment shown by the algorithm would be dangerous and vice-versa.

Typically the methodology of the present invention requires at least one image capturing device (701) such as camera or surveillance camera for acquiring reference image and scene image of said object of interest. Said image capturing device (701) can be a single 2D camera. However, it shall be appreciated that a plurality of cameras or 3D camera can also be used as long as image of the object of interest is obtained. Alternatively, said WMSN can comprises of Intelmote2 (IPR2400) WSN modules, multimedia sensor board (IMB400) and IIB2400 interface board which are integrated with said image capturing device (701) to increase the efficiency of the system and is mounted anywhere as long as the image for said object of interest is able to be acquired, for example on the transmission pole at the other end wherefrom the image is captured. Said Intelmote2 uses IEEE 802.15.4 standard that is regarded as Low-Rate Personal Area Network (LR-PAN) and is called ZigBee. Said multimedia sensor incorporates an imaging sensor that is capable of acquiring RGB colour images having an appropriate resolution, for example 640 x 480 pixels. Said IPR2400 incorporates CC2420 radio module integrated with λ/2 dipole antenna having a transmission power between -24 dBm and 0 dB and uses a low bit-rate (250kbps) to communicate wirelessly with other WSN modules. For high bit-rate, Xbee module is integrated with Intelmote2 to provide a long range (upto 7km) communication.

Wireless Sensor Networks (WSN) are installed along the transmission poles right-of-ways (ROWs) and can be wirelessly interlinked to one another forming a monitoring network (i.e. WMSN), as shown in FIG. 9. Said WSNs are integrated to said image capturing device (701) so that the acquired image data is transmitted to monitoring station by communicating with one another in a Multi-hop manner. The monitoring station is provided with at least one computer with wireless Gateway which includes Ethernet and 3G interfaces like CDMA/ GSM/ GPRS, said computer can further be connected to internet so that multiple users can observe the monitored vegetation encroachment conditions. The band for communication that will be used is 2.4GHz as it is unregistered globally. Energy source such as solar battery is used to power the WSN and said image capturing device integrated with IMB400 on the poles right-of -ways. Said image capturing device and WSN are sealed in a plastic type coating and a windshield to avoid any problem from electromagnetic (EM) radiations of transmission lines and bad weather. At the monitoring station, at least one electronic processing means is provided for processing and analysing said acquired image. Said images are passed through said electronic processing means which comprises of a graphical user interface (GUI) embedded with the algorithm for processing of said images as well as online monitoring of transmission lines against dangerous encroachments within and outside right-of-ways. At the monitoring station, images would be processed to identify the weather conditions using pattern recognition methods which includes PCA and "Dark Channel Prior Model". After that image processing based object tracking is used to detect the vegetations and a depth from triangulation based method is used that identifies actual 3D coordinates of the encroached vegetations. That helps in locating the height of trees, distance between trees and poles and distance between dangerous trees and transmission lines outside right-of-ways. Alternatively, said electronic processing means also can be positioned onsite to process and analyse said acquired image before being transmitted to monitoring station by WMSN.

After receiving the information at the monitoring station, the authorities or the operators at the monitoring station will order the trimmers to cut-off the vegetation where ever it is required. Thus this method as compared to other airborne techniques will prove to be cost effective, lesser time consuming and more accurate. While the method for vegetation encroachment is applied to transmission line, the present invention is not restricted to this but it shall also be understood that this also can be applied to power line pole, distribution line pole, communication cable pole and construction structures. While the preferred embodiment of the present invention and its advantages has been disclosed in the above Detailed Description, the invention is not limited there to but only by the scope of the appended claim.