Process and system for tracking ground movement
The present invention employs video camera-equipped, unmanned aerial vehicles (UAVs) to collect video imagery from a predefined study area on the ground, where vehicles are performing trips. The video imagery is then processed using machine vision to extract vehicle trajectory information, which is then organized and manipulated for presentation in 5 a user-facing display when required. In contrast with the state of the art data procurement technologies, the resulting data features continuity in both time and space, reflects 100% of the trips and has a significantly lower cost.
1 . A process for tracking ground movement, comprising the steps of: simultaneously deploying two or more Unmanned Aerial Vehicles (UAVs) to collect ground traffic video imagery, a first UAV equipped with a first camera that collects ground video imagery from a first polygonal coverage area on the ground during a survey period and a second UAV equipped with a second camera that collects ground video imagery from a second polygonal coverage area during the same survey period, wherein the first and second cameras collect video imagery simultaneously during the survey period and the first and second polygonal coverage areas share an edge; combining the imagery of the first and second cameras to form a continuous ground video imagery mosaic within which the movement of one or more ground vehicles and/or one or more pedestrians can be tracked with continuity in both time and space.
2 . The process according to claim 1 wherein each polygonal coverage area on the ground is defined by a portion of the ground that remains visible in the corresponding camera video imagery in spite of the continuous movement of the UAV to which said camera is attached.
3 . The process according to claim 2 , wherein variable video imagery distortion resulting from the constant movement of the UAV to which the camera is attached is attenuated by software corrections timed by the rotational frequency of the UAV's circular flight path.
4 . The process according to claim 1 , wherein during the survey period each UAV flies a circular flight path of reduced radius that is preferably shorter than 50 meters long.
5 . The process according to claim 1 , where the imagery collected by each UAV camera has a corresponding ground sampling distance, wherein the UAV flies at higher altitude and said ground sampling distance is kept unchanged by equipping the UAV camera with a telescopic lens that reduces the camera's aperture angle.
6 . The process according to claim 1 , where the imagery collected by each UAV camera has a corresponding ground sampling distance, wherein the UAV flight altitude, the selected camera's video frame resolution and aperture angle are combined to achieve a Ground Sampling Distance no higher than 500 mm/px in the continuous video imagery mosaic.
7 . The process according to claim 1 , wherein the camera in each UAV is mounted on a motorized gimbal that keeps the camera pointed vertically downwards and prevents its rotation in the vertical axis by successive adjustments performed via actuation of the gimbal's motors, regardless of the position and orientation of the UAV during its survey flight.
8 . The process according to claim 1 , wherein the movement of the one or more vehicles and/or pedestrians within the video imagery mosaic is tracked during the survey period using machine vision to track the successive vehicular and/or pedestrian positions within the video imagery frames via pixel differentiation over time.
9 . The process according to claim 1 , wherein the positional reference of a central point of each polygonal coverage area that makes up the video imagery mosaic is used to create trajectory files that document the movement of each individual vehicle and/or pedestrian tracked within the video imagery mosaic as a series of consecutive sets of GPS coordinates, such that vehicle and/or pedestrian trajectories can be matched to a subjacent street network within a study area.
10 . The process according to claim 9 , wherein the vehicle and/or pedestrian trajectory files are processed to associate vehicle classification to each vehicle and/or pedestrian based on its unique trajectory file information.
11 . The process according to claim 10 , wherein the vehicle and/or pedestrian trajectories and vehicle and/or pedestrian classifications are processed to a customizable software display user interface, configured to permit a user to select the data that gets presented, such as by manipulation of the vehicle trajectory files.
12 . The process according to claim 1 , wherein the UAV is an untethered, fixed-wing aircraft.
13 . The process according to claim 1 , wherein the first or second camera comprises an array of two or more cameras.
14 . The process according to claim 1 , wherein a plurality of UAVs fly in relays to accommodate any combination of study area size and analysis period.
15 . A system for tracking ground movement of at least one vehicle and/or pedestrian on the ground, comprising a computer system comprising: a processor and a non-transitory computer readable medium storing instructions that, when executed by the processor, cause the processor to perform a method comprising the steps of:
a. accessing aerial ground video imagery taken simultaneously above a first polygonal coverage area and a second polygonal coverage area, wherein the first and second polygonal coverage areas share an edge;
b. combining the ground video imagery of the first and second polygonal coverage areas to form a continuous ground video imagery mosaic; and
c. tracking the movement of the at least one ground vehicle and/or pedestrian within the first and second polygonal coverage areas, with continuity in both time and space.
16 . The system according to claim 15 wherein the tracking of the movement within said continuous ground video imagery mosaic is performed via pixel differentiation over time.