System for inspecting and assessing the condition of the pavement of a roadway segment

1 . A system for inspecting and assessing a condition of a pavement of a roadway segment, the system comprises:

at least one drone that comprises an imaging device configured to capture image data having pixels representing a surface of the pavement, a geopositioning module configured to determine geoposition data of the at least one drone, and one or more sensors configured to acquire operational data of the at least one drone, operational data of the imaging device or environmental data; and

an AI-based computer-vision image processing unit configured to execute a first process and a second process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the imaging device, the geopositioning module, and the one or more sensors are configured to operate simultaneously to enable correlation of the image data, the geoposition data of the at least one drone, and one or more of the operational data of the at least one drone, the operational data of the imaging device, and the environmental data, thereby enabling a georeferencing of the pixels within the captured image data;

wherein the first process comprises extracting one or more image data portions from the captured image data through a first screening process, each image data portion having less than a full spatial extent of the captured image data and comprising one or more sets of pixels, each corresponding to a candidate pavement defect; independently processing each image data portion independently by a first plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the first plurality of AI-based computer-vision algorithms treats each image data portion as a separate input and generates each generating a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the first plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated by the system as a pavement defect; measuring each pavement defect; geopositioning each pavement defect by georeferencing its respective set of pixels; and determining the condition measure of the pavement using the measured and geopositioned pavement defects; and

wherein the second process comprises generating an orthophoto from the captured image data, wherein the orthophoto has georeferenced pixels representing the surface of the pavement; extracting one or more orthophoto portions from the generated orthophoto through a second screening process, each orthophoto portion having less than the full spatial extent of the generated orthophoto and comprising one or more sets of georeferenced pixels, each corresponding to a candidate pavement defect; independently processing each orthophoto portion independently by a second plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the second plurality of AI-based computer-vision algorithms treats each orthophoto portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming the presence of each candidate pavement defect based on concordant pavement defect detections generated by the second plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated by the system as a pavement defect, which is already geopositioned by its respective set of georeferenced pixels; measuring each pavement defect; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.

2 . The system of claim 1 , wherein the at least one drone follows only two flight trajectories, each flown once, a first extending along the entire roadway segment in one A direction and a second extending along the entire roadway segment in an opposite direction.

3 . The system of claim 2 , wherein the two flight trajectories extend on opposite sides of a centerline of the roadway segment.

4 . The system of claim 1 , wherein the geopositioning and measurement of the pavement defects are performed with an accuracy of about 0.4 inches or less, and wherein the condition measure is a PCI.

5 . The system of claim 4 , wherein the surface of the pavement of the roadway segment is inspected in its entirety and its condition assessed, such that the PCI is based on a complete inspection rather than a statistical sampling.

6 . The system of claim 1 , wherein the at least one drone is configured to follow flight trajectories having controlled altitude, speed and turning radii to reduce or eliminate motion blur in the captured image data.

7 . The system of claim 1 , wherein the first and second pluralities of AI-based computer-vision algorithms comprises algorithms belonging to different families of AI-based computer-vision algorithms.

8 . The system of claim 1 , wherein the geopositioning module uses RTK satellite data to determine the geoposition data of the at least one drone.

9 . The system of claim 1 , wherein the roadway segment comprises flexible pavement, rigid pavement, or a combination thereof, and the condition measure is determined separately for each pavement type.

10 . The system of claim 1 , wherein the system provides an inventory of the measured and geopositioned pavement defects as part of a GIS map layer.

11 . A method for inspecting and assessing a condition of a pavement of a roadway segment, the method comprises:

flying at least one drone along the roadway segment;

capturing image data having pixels representing a surface of the pavement using the at least one drone; and

executing a first process and a second process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the first process comprises extracting one or more image data portions from the captured image data through a first screening process, each image data portion having less than a full spatial extent of the captured image data and comprising one or more sets of pixels, each corresponding to a candidate pavement defect; independently processing each image data portion independently by a first plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the first plurality of AI-based computer-vision algorithms treats each image data portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the first plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated as a pavement defect; measuring each pavement defect; geopositioning each pavement defect by georeferencing its respective set of pixels based on geoposition data of the at least one drone; and determining the condition measure of the pavement using the measured and geopositioned pavement defects; and

wherein the second process comprises generating an orthophoto from the captured image data using geoposition data of the at least one drone, wherein the orthophoto has georeferenced pixels representing the surface of the pavement; extracting one or more orthophoto portions from the generated orthophoto through a second screening process, each orthophoto portion having less than the full spatial extent of the generated orthophoto and comprising one or more sets of georeferenced pixels, each corresponding to a candidate pavement defect; processing each orthophoto portion independently by a second plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the second plurality of AI-based computer-vision algorithms treats each orthophoto portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming the presence of each candidate pavement defect based on concordant pavement defect detections generated by the second plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated as a pavement defect, which is already geopositioned by its respective set of georeferenced pixels; measuring each pavement defect; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.

12 . The method of claim 11 , wherein the at least one drone follows only two flight trajectories, each flown once, a first extending along the entire roadway segment in one direction and a second extending along the entire roadway segment in an opposite direction.

13 . The method of claim 12 , wherein the two flight trajectories extend on opposite sides of a centerline of the roadway segment.

14 . The method of claim 11 , wherein the geopositioning and measurement of the pavement defects are performed with an accuracy of about 0.4 inches or less, and wherein the condition measure is a PCI.

15 . The method of claim 14 , wherein the surface of the pavement of the roadway segment is inspected in its entirety and its condition assessed, such that the PCI is based on a complete inspection rather than a statistical sampling.

16 . The method of claim 11 , wherein the at least one drone is configured to follow flight trajectories having controlled altitude, speed and turning radii to reduce or eliminate motion blur in the captured image data.

17 . The method of claim 11 , wherein the first and second pluralities of AI-based computer-vision algorithms comprises algorithms belonging to different families of AI-based computer-vision algorithms.

18 . The method of claim 11 , wherein RTK satellite data is used to determine the geoposition data of the at least one drone.

19 . The method of claim 11 , wherein the roadway segment comprises flexible pavement, rigid pavement, or a combination thereof, and the condition measure is determined separately for each pavement type.

20 . The method of claim 11 , wherein the system provides an inventory of the measured and geopositioned pavement defects as part of a GIS map layer.

21 . A system for inspecting and assessing a condition of a pavement of a roadway segment, the system comprises:

at least one drone that comprises an imaging device configured to capture image data having pixels representing a surface of the pavement, a geopositioning module configured to determine geoposition data of the at least one drone, and one or more sensors configured to acquire operational data of the at least one drone, operational data of the imaging device or environmental data; and

an AI-based computer-vision image processing unit configured to execute a process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the imaging device, the geopositioning module, and the one or more sensors are configured to operate simultaneously to enable correlation of the image data, the geoposition data of the at least one drone, and one or more of the operational data of the at least one drone, the operational data of the imaging device, and the environmental data, thereby enabling a georeferencing of the pixels within the captured image data; and

wherein the process comprises extracting one or more image data portions from the captured image data through a screening process, each image data portion having less than a full spatial extent of the captured image data and comprising one or more sets of pixels, each corresponding to a candidate pavement defect; independently processing each image data portion by a plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the plurality of AI-based computer-vision algorithms treats each image data portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated by the system as a pavement defect; measuring each pavement defect; geopositioning each pavement defect by georeferencing its respective set of pixels; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.

22 . A system for inspecting and assessing a condition of a pavement of a roadway segment, the system comprises:

at least one drone that comprises an imaging device configured to capture image data having pixels representing a surface of the pavement, a geopositioning module configured to determine geoposition data of the at least one drone, and one or more sensors configured to acquire operational data of the at least one drone, operational data of the imaging device or environmental data; and

an AI-based computer-vision image processing unit configured to execute a process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the imaging device, the geopositioning module, and the one or more sensors are configured to operate simultaneously to enable correlation of the image data, the geoposition data of the at least one drone, and one or more of the operational data of the at least one drone, the operational data of the imaging device, and the environmental data, thereby enabling a georeferencing of the pixels within the captured image data; and

wherein the process comprises generating an orthophoto from the captured image data, wherein the orthophoto has georeferenced pixels representing the surface of the pavement; extracting one or more orthophoto portions from the generated orthophoto through a screening process, each orthophoto portion having less than a full spatial extent of the generated orthophoto and comprising one or more sets of georeferenced pixels, each corresponding to a candidate pavement defect; independently processing each orthophoto portion by a plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the plurality of AI-based computer-vision algorithms treats each orthophoto portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated by the system as a pavement defect, which is already geopositioned by its respective set of georeferenced pixels; measuring each pavement defect; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.

23 . A method for inspecting and assessing a condition of a pavement of a roadway segment, the method comprises:

flying at least one drone along the roadway segment;

capturing image data having pixels representing a surface of the pavement using the at least one drone; and

executing a process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the process comprises extracting one or more image data portions from the captured image data through a screening process, each image data portion having less than a full spatial extent of the captured image data and comprising one or more sets of pixels, each corresponding to a candidate pavement defect; independently processing each image data portion by a plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the plurality of AI-based computer-vision algorithms treats each image data portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated as a pavement defect; measuring each pavement defect; geopositioning each pavement defect by georeferencing its respective set of pixels based on geoposition data of the at least one drone; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.

24 . A method for inspecting and assessing a condition of a pavement of a roadway segment, the method comprises:

flying at least one drone along the roadway segment;

capturing image data having pixels representing a surface of the pavement using the at least one drone; and

executing a process for detecting, geopositioning and measuring pavement defects from the captured image data, and determining a condition measure of the pavement;

wherein the process comprises generating an orthophoto from the captured image data using geoposition data of the at least one drone, wherein the orthophoto has georeferenced pixels representing the surface of the pavement; extracting one or more orthophoto portions from the generated orthophoto through a screening process, each orthophoto portion having less than a full spatial extent of the generated orthophoto and comprising one or more sets of georeferenced pixels, each corresponding to a candidate pavement defect; independently processing each orthophoto portion by a plurality of AI-based computer-vision algorithms, wherein each AI-based computer-vision algorithm of the plurality of AI-based computer-vision algorithms treats each orthophoto portion as a separate input and generates a pavement defect detection for each candidate pavement defect; confirming a presence of each candidate pavement defect based on concordant pavement defect detections generated by the plurality of AI-based computer-vision algorithms, wherein each confirmed candidate pavement defect is treated as a pavement defect, which is already geopositioned by its respective set of georeferenced pixels; measuring each pavement defect; and determining the condition measure of the pavement using the measured and geopositioned pavement defects.