Abstract: This disclosure generally relates to method and system for asset inspection using unmanned aerial vehicles. Regular inspection of assets in any environment is time consuming and involves physical risk and injury to the property. The method includes receiving a plurality of input images of an asset using unmanned aerial vehicle and generates a 3D point cloud based on overlapped images identified from the plurality of input images of the asset being inspected. Further, anomalies are detected and defects are observed in the one or more objects of interests based on the change observed in each 2D image frame from the plurality 2D image frames in a specific view with an identical objects of interest of the 2D image frame. A pyramidal view map is created for the plurality 2D image frames by mapping each pixel of the plurality 2D image frames with the plurality of input images.

Abstract: Detection of textual phrases in a non-horizontal orientation at a scene is a target problem. This disclosure relates to a processor implemented method to detect a text from multimedia content captured at a scene. An input original image is processed by a trained model to obtain individual character with bounding box on the original image. The original image is positioned by a gradient to obtain a rotated image if number of detected characters is not equal to number of expected characters on the original image. At least one missing character bounding box on the original image and on the rotated image are estimated to construct a horizontal text image if number of detected characters is not equal to number of expected characters on the rotated image. At least one missing character in the estimated bounding box is detected by at least one text returned from an optical character reader.

Abstract: This disclosure relates generally to method and system for tracking motion of subjects in three dimensional space. The method includes receiving a video of the environment using a scene capturing device positioned in the environment. A motion intensity of subjects from the plurality of image frames are detected for segregating the motion of subjects present in each image frame from the plurality of image frames into a plurality of categories. Further, a three dimensional (3D) scene from the plurality of image frames are constructed using the multi focused view based depth calculation technique. The subjects are tracked based on the position in three dimensional (3D) scene categorized under the significant motion category. The proposed disclosure provides efficiency in tracking the new entry of subjects in the environment for adjusting the focus of observer.

Abstract: This disclosure generally relates to method and system for asset inspection using unmanned aerial vehicles. Regular inspection of assets in any environment is time consuming and involves physical risk and injury to the property. The method includes receiving a plurality of input images of an asset using unmanned aerial vehicle and generates a 3D point cloud based on overlapped images identified from the plurality of input images of the asset being inspected. Further, anomalies are detected and defects are observed in the one or more objects of interests based on the change observed in each 2D image frame from the plurality 2D image frames in a specific view with an identical objects of interest of the 2D image frame. A pyramidal view map is created for the plurality 2D image frames by mapping each pixel of the plurality 2D image frames with the plurality of input images.

Abstract: Vehicles are rigid/multi-dimensional objects capable of moving through space which do not undergo structural deformation. Existing systems/techniques analyse videos (e.g., traffic videos) for tracking moving objects either manually or through crowd sourcing and thus are less accurate due to different variations (e.g., non-linear scaling, visibility, occlusion, etc.) caused due to perspective. Present disclosure enable real-time video analytics wherein background from images in de-noised motion streams is filtered to obtain moving objects of interest. Motion map is generated to identify moving objects along-with depth information.

Abstract: This disclosure relates generally to method and system for tracking motion of subjects in three dimensional space. The method includes receiving a video of the environment using a scene capturing device positioned in the environment. A motion intensity of subjects from the plurality of image frames are detected for segregating the motion of subjects present in each image frame from the plurality of image frames into a plurality of categories. Further, a three dimensional (3D) scene from the plurality of image frames are constructed using the multi focused view based depth calculation technique. The subjects are tracked based on the position in three dimensional (3D) scene categorized under the significant motion category. The proposed disclosure provides efficiency in tracking the new entry of subjects in the environment for adjusting the focus of observer.

Abstract: Current technologies detect the alertness of a user using frontal face. However, there are no techniques to track the position of face from a side angle. A method and system for face position tracking of a user and alerting the user is disclosed. The method includes capturing images using a monocular camera and identifying a user's face from the images and storing the user's face as reference face. The user's face is divided into two regions and one or more corner points are identified. A centroid is created in the two regions and the corner points are joined using virtual lines to create a dual flexible spider model. One or more new corner points are created on fulfillment of pre-defined conditions. Angle and magnitude of flow vectors is determined using one or more new corner points and alerts are given to the user based on the angle and magnitude.

Abstract: Vehicles are rigid/multi-dimensional objects capable of moving through space which do not undergo structural deformation. Existing systems/techniques analyse videos (e.g., traffic videos) for tracking moving objects either manually or through crowd sourcing and thus are less accurate due to different variations (e.g., non-linear scaling, visibility, occlusion, etc.) caused due to perspective. Present disclosure enable real-time video analytics wherein background from images in de-noised motion streams is filtered to obtain moving objects of interest. Motion map is generated to identify moving objects along-with depth information.

Abstract: Current technologies detect the alertness of a user using frontal face. However, there are no techniques to track the position of face from a side angle. A method and system for face position tracking of a user and alerting the user is disclosed. The method includes capturing images using a monocular camera and identifying a user's face from the images and storing the user's face as reference face. The user's face is divided into two regions and one or more corner points are identified. A centroid is created in the two regions and the corner points are joined using virtual lines to create a dual flexible spider model. One or more new corner points are created on fulfillment of pre-defined conditions. Angle and magnitude of flow vectors is determined using one or more new corner points and alerts are given to the user based on the angle and magnitude.

Abstract: This disclosure relates generally to lane detection, and more particularly to method and systems for lane detection. In one embodiment, a method for lane detection is includes generating an edge image including inbound edges of lane-markings corresponding to a road image. Based on an adaptive thresholding of intensity of lane-markings from image strips selected from the image, a plurality of lane masks are extracted. A plurality of valid lane masks are selected from the plurality of lane masks based on a structure and intensity information of the plurality of valid lane masks in the image strips. An overlap between the inbound edges and the valid lane masks is determined to identify a plurality of edge segments of the pair of inbound edges belonging to the lane-markings. A set of valid edge segments of the pair of inbound edges are obtained based on the curvature information of the edge segments.

Abstract: Systems and methods for ground truth determination in lane departure warnings are provided. The methods include receiving time slice images of a lane captured at different time frames from an image capturing unit. The intensity profiles of these time slice images are determined and these intensity profiles of the images are smoothened to obtain a smoothened histogram. A threshold value of these time sliced images are determined and further refined to extract the lane marking of a lane. The lane extended to multiple rows of a lane to determine the ground truth value used for validating a lane departure warning.

Abstract: A solar receiver, designed to use a heat transfer medium, includes a plurality of panels. Each panel is arranged and configured to enable the heat transfer medium to flow in at least one flow direction, one flow direction defining a pass, to obtain unique mass flux in each pass to optimize the heat flux capability of the pass while minimizing pressure drop across the selected passes of the heat transfer surface. A method thereof is also provided.

Abstract: A solar thermal power system includes a solar receiver for heating thermal energy storage fluid to be stored and utilized from a thermal energy storage arrangement having hot and cold storage tanks. The system includes a steam generator arrangement, which utilizes the heat of the thermal energy storage fluid to produce steam to run a turbine. The arrangement includes a bypass line configured to bypass the hot storage tank from the steam generator arrangement, and to supply the hot thermal energy storage fluid from the solar receiver directly to the steam generator arrangement, during day times, when the solar receiver the steam generator arrangement are both in operating mode, thereby recovering stored potential energy available in the down corner hot thermal energy storage fluid from the solar receiver.

Abstract: Methods, systems, and computer program products for improving the performance of a raster image processor. Smaller objects are identified among a group of larger objects with respect to a job processed via a raster image processor. The smaller objects are merged with one or more larger objects among the group of larger objects. The smaller objects that are merged with the larger object are treated as a single object without losing the perceptual quality of the job and while reducing memory requirements to thereby enhance productivity during processing of the job via the raster image processor. RIP performance improvement results by pre-flattening complicatedly designed backgrounds with multiple objects of significantly low relative occupancy.

Abstract: Disclosed is a method and video surveillance system for detecting fraud. The system may receive images. The system may convert the images into binary images. Subsequently, the system may identify a reference frame amongst the binary images. The system may trigger face detection upon detecting motion within the binary images. The system may detect the motion based on a relative white pixel count of the binary image with respect to a reference image. The system may reduce a Region of Interest (RoI) window within the binary images based on a highest density of white pixel count. The system may determine whether a face is detected within the reduced RoI window. The system may determine whether mouth is detected within a lower half of the face when the face is detected. The system may detect the fraud when the face is not detected or the mouth is not detected.

Abstract: Methods, systems, and computer program products for improving the performance of a raster image processor. Smaller objects are identified among a group of larger objects with respect to a job processed via a raster image processor. The smaller objects are merged with one or more larger objects among the group of larger objects. The smaller objects that are merged with the larger object are treated as a single object without losing the perceptual quality of the job and while reducing memory requirements to thereby enhance productivity during processing of the job via the raster image processor. RIP performance improvement results by pre-flattening complicatedly designed backgrounds with multiple objects of significantly low relative occupancy.

Abstract: This disclosure relates generally to lane detection, and more particularly to method and systems for lane detection. In one embodiment, a method for lane detection is includes generating an edge image including inbound edges of lane-markings corresponding to a road image. Based on an adaptive thresholding of intensity of lane-markings from image strips selected from the image, a plurality of lane masks are extracted. A plurality of valid lane masks are selected from the plurality of lane masks based on a structure and intensity information of the plurality of valid lane masks in the image strips. An overlap between the inbound edges and the valid lane masks is determined to identify a plurality of edge segments of the pair of inbound edges belonging to the lane-markings. A set of valid edge segments of the pair of inbound edges are obtained based on the curvature information of the edge segments.

Abstract: Methods, systems, and computer-program products for optimizing SRE (Super Resolution Encoding) patterns. A hierarchical self-organizing pattern map (HSOPM) of SRE patterns can be derived, which illustrates interrelationships between consecutive SRE patterns. Such a hierarchical self-organizing map provides a first level of hierarchy, a second level of hierarchy, etc. Different weights can be assigned to different synthesis of traversal (SoT) according to the second level of hierarchy. The likelihood of the SRE patterns can then be calculated based on a fitness of continuity and the different weights, so as to subsequently select and encode an allowed number of the SRE patterns while replacing other patterns with a lower likelihood value with an immediate root and thereby adaptively optimize any number of the SRE patterns with respect to any number of values.

Abstract: The disclosed embodiments illustrate methods and systems for encoding an image. The method includes identifying one or more objects, having an associated first tag value, in the image. The first tag value is deterministic of at least a type of the one or more objects. The number of the one or more objects in the image is less than a predetermined number of objects. The method further includes assigning a second tag value to each pixel in the image to create an encoded image. The second tag value is assigned based on the type of object represented by each pixel. The size of the second tag value is less than the size of the first tag value. The method further includes defining a header field for the encoded image. The header field includes the first tag value associated with each of the one or more objects.

Abstract: Systems and methods for ground truth determination in lane departure warnings are provided. The methods include receiving time slice images of a lane captured at different time frames from an image capturing unit. The intensity profiles of these time slice images are determined and these intensity profiles of the images are smoothened to obtain a smoothened histogram. A threshold value of these time sliced images are determined and further refined to extract the lane marking of a lane. The lane extended to multiple rows of a lane to determine the ground truth value used for validating a lane departure warning.