Methods and systems for shape based image analysis for detecting linear objects

A system for shape based image analysis of linear objects, the system comprising:

• one or more processors;
• and
• one or more internal data storage devices operatively coupled to the one or more processors for storing instructions configured for execution by the one or more processors, the instructions being comprised in:
• receive at least one image of the linear objects, wherein the at least one image of the linear object corresponds to a linear object of interest, wherein the linear object is an object comprising at least a portion of the object with a linear and a finite length in the at least one image;
• a blob extractor configured to extract homogeneous blobs having polygonal contours from the at least one image of the linear objects of interest;
• a trapezium classifier configured to classify the extracted polygons contours by identifying a quadrilateral resembling contour based on shape of each blob of the homogeneous blobs, wherein the classified quadrilateral-resembling contour is identified as candidate trapeziums having at least one pair of parallel edges;
• a clustering module configured to cluster the candidate trapeziums based on spatial orientation and alignment of medial axes of the candidate trapezium and sorting the candidate trapeziums based on height of the candidate trapeziums to obtain longer trapeziums;
• and
• an analyzing and detecting module configured to detect the linear objects of interest from the longer trapeziums based on context information, wherein the context information is a cross arm associated with each of the linear objects of interest, wherein the cross arm is detected based on the extracted polygonal contours, height of the longer trapeziums and angle of intersection of the cross arm with the detected linear objects of interest;
• and
• classify the detected linear objects of interest as damaged based on at least one of height of the detected linear objects and inclination of the detected linear objects with respect to ground.

A computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:

• receive at least one image of the linear objects, wherein the at least one image of the linear object corresponds to a linear object of interest, wherein the linear object is an object comprising at least a portion of the object with a linear and a finite length in the at least one image;
• extract homogeneous blobs from the at least one image using a blob extractor having polygonal contours;
• classify the extracted polygons contours by identifying a quadrilateral resembling contour based on shape of each blob of the homogeneous blobs, wherein the classified quadrilateral-resembling contour is identified as candidate trapeziums having at least one pair of parallel edges;
• clustering, the candidate trapeziums based on spatial orientation and alignment of medial axes of the candidate trapezium and sorting the candidate trapeziums based on height of the candidate trapeziums to obtain longer trapeziums;
• detect the linear objects of interest from the longer trapeziums based on context information wherein the context information is a cross arm associated with each of the linear objects of interest, wherein the cross arm is detected based on the extracted polygonal contours, height of the longer trapeziums and angle of intersection of the cross arm with the detected linear objects of interest;
• and
• classifying the detected linear objects of interest as damaged based on at least one of height of the detected linear objects and inclination of the detected linear objects with respect to ground.