CROWD DETECTION, ANALYSIS, AND CATEGORIZATION

• US 20180253606A1
• Filed: 09/28/2017
• Published: 09/06/2018
• Est. Priority Date: 03/03/2017
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Dhua, Silpi, Omanwar, Anil M., Sett, Sujoy, Waykos, Pradip A.
• Application Number:
  US15/718,113
• Publication Number:
  US 20180253606A1
• Time in Patent Office: Days
• CPC Class Codes: B64U 2101/30 for imaging, photography or videography
• CPC Class Codes: B64U 2201/10 autonomous, i.e. by navigating independently from ground or air stations, e.g. by using inertial navigation systems [INS]
• CPC Class Codes: G05D 1/0094 involving pointing a payload, e.g. camera, weapon, sensor, towards a fixed or moving target
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06Q 30/0241 Advertisements
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 20/17 taken from planes or by drones
• CPC Class Codes: G06V 20/53 Recognition of crowd images, e.g. recognition of crowd congestion
• CPC Class Codes: G08G 5/0069 specially adapted for an unmanned aircraft

Abstract

A method, computer system, and a computer program product for analyzing a crowd using a plurality of images captured by an aerial drone is provided. The present invention may include determining a geographic area associated with the crowd. The present invention may also include partitioning the determined geographic area into a plurality of zones. The present invention may then include determining a flight path covering each zone within the plurality of zones. The present invention may further include receiving the plurality of images from the aerial drone. The present invention may also include analyzing the received plurality of images to identify a plurality of individuals associated with the crowd. The present invention may then include predicting a plurality of crowd characteristics based on the analyzed plurality of images. The present invention may further include performing an action in response to the predicted plurality of crowd characteristics.

Claims

1. A method for analyzing a crowd using a plurality of images captured by an aerial drone, the method comprising:

determining a geographic area associated with the crowd;

partitioning the determined geographic area into a plurality of zones;

determining a flight path covering each zone within the plurality of zones;

sending the determined flight path to the aerial drone;

flying, by the aerial drone, along the sent flight path;

generating, by the aerial drone, the plurality of images, and a plurality of drone position data that comprises a drone location, a drone orientation, and a photographed location;

receiving the plurality of images and the plurality of drone position data from the aerial drone;

analyzing the received plurality of images to identify a plurality of individuals associated with the crowd, wherein analyzing the received plurality of images to identify the plurality of individuals associated with the crowd further comprises cropping each image within the received plurality of images to create an image partition for each identified individual within the identified plurality of individuals;

determining an individual position for each identified individual within the identified plurality of individuals based on the analyzed plurality of images and the received plurality of drone position data;

determining an individual movement vector for each identified individual within the identified plurality of individuals based on tracking changes in the determined individual position associated with the individual as captured within the analyzed plurality of images;

predicting a plurality of individual characteristics for each identified individual based on the analyzed plurality of images, wherein predicting the plurality of individual characteristics comprises using a pre-trained machine learning model to process the analyzed plurality of images and predict the plurality of individual characteristics, and wherein the predicted plurality of individual characteristics includes a plurality of demographic characteristics and a plurality of physical characteristics;

predicting a plurality of crowd characteristics based on the predicted plurality of individual characteristics for the plurality of identified individuals, wherein the predicted plurality of crowd characteristics includes an aggregate motion characteristic determined based on the determined individual movement vector for each identified individual and a rate of entry and a rate of exit corresponding with each zone within the plurality of zones, and wherein the predicted plurality of crowd characteristics includes a crowd concentration characteristic based on the determined individual position for each identified individual; and

determining a placement and a duration for dynamic advertising to distribute to the crowd based on the predicted plurality of crowd characteristics.