UAS Regulations and Usage. July 14, 2016 Presenter: David Day, CP, GISP Keystone Aerial Surveys, Inc.

Size: px

Start display at page:

Download "UAS Regulations and Usage. July 14, 2016 Presenter: David Day, CP, GISP Keystone Aerial Surveys, Inc."

Beryl Miller
6 years ago
Views:

2 UAS Regulations and Usage July 14, 2016 Presenter: David Day, CP, GISP Keystone Aerial Surveys, Inc.

3 Keystone Aerial Surveys, Inc. Based in Philadelphia, PA with offices in Los Angeles, Tucson and Tyler, TX Founded in 1963 to provide remote sensing throughout the US 21 manned survey aircraft 22 mapping sensors including 8 UltraCam digital cameras and 2 Optech LiDAR systems (including new Galaxy delivered June 14 th!) Provide post production services from large format digital, LiDAR and UAS imagery

Keystone UAS UAS approved: Altavian Nova F6500 (fixed wing), R8400 (Rotorcraft) DJI Phantom 3, Inspire 1 Pro and DJI S1000 (Rotorcraft) 3D Robotics X8-M (Rotorcraft)

4 Keystone UAS UAS approved: Altavian Nova F6500 (fixed wing), R8400 (Rotorcraft) DJI Phantom 3, Inspire 1 Pro and DJI S1000 (Rotorcraft) 3D Robotics X8-M (Rotorcraft) SteadiDrone Mavrik X4 and Mavrik X8 (Rotorcraft) Keystone is approved for all types of mapping, remote sensing and imagery except closed set television and film capture

5 My Bio Started with Keystone in 1995 as an aerial photographer Developed Information Technology and Digital Post Processing departments Have been hands on with imagery post processing software for nearly 10 years Application Development Certification, Certified Photogrammetrist and Geographic Information Systems Professional (GISP) Currently Executive Vice President and Director of UAS Division Oversee all manned and unmanned flight operations, camera systems and post production Authored or Co-Authored 7 research papers, many with mentor Dr. Ricardo Passini, BAE systems.

7 Current Regulations Must have a 333 exemption to legally fly UAS for commercial purposes Exemptions are from the FAA for specific airframes and organizations\persons Hobbyists are not required to obtain exemptions Exemption holders must have a Certificate of Authorization (COA) to fly their exempted aircraft All flights involving commercial operations must be performed under 333 and COA or they are ILLEGAL!

8 Current Regulations 333 Key Rules UAV must be under 55 lbs. and speed not to exceed 100 mph UAV must be operated at 400 feet AGL or below UAV must be operated within visual line of sight of pilot and one observer (cannot daisy-chain observers/pilots) Pilot must hold a commercial, private, recreational or sport pilot certificate Operations cannot be at night Operations must be in VFR conditions (500 ft. below clouds)

Current Regulations 333 Key Rules UAV may not operate within 5 miles of a public airport without permission All operations must be 500 feet from

9 Current Regulations 333 Key Rules UAV may not operate within 5 miles of a public airport without permission All operations must be 500 feet from buildings, vessels, persons and vehicles not participating in the operation Exemption holders receive a blanket COA and access to the Civil COA request website

10 Current Regulations Blanket COA Rules Updated on March 29 th, 2016 with new COAs released on April 4 th, 2016 allowing operations up to 400 Monthly reporting of all activities and location of UAS activities File a NOTAM for each operation Must be 5 nm from public airport with a tower, 3 nm from most other public airports and 2 nm from public heliport, gliderport or seaport

11 New Regulations as of August 29th FAA Release new Part 107 on June 21 st setting permanent rules for use of UAS for commercial purposes Operations must not be over people Operations must be at 400 or below Except when flying near a building or tower must maintain 400 laterally and vertically, but can exceed the altitude restriction Can Operate from a moving vehicle (if in a remote area) Can only operate in Class G airspace

12 Left: Airspace in NJ Below: Class C airspace with Class E airspace going to ground. The dashed purple line is off limits to drones but does not show on most airspace databases

13 New Regulations as of August 29th

14 Keystone developed UAS airspace KMZ with surface airspace and restricted areas where we cannot fly

15 Locations prevented from flying under 333 exemption. New rules will be much more friendly to UAS

16 New Regulations as of August 29th Pilot Must have a Remote Pilot Certification with Small UAS Rating Must pass a written test at an FAA approved testing facility Must be 16 or older Must pass an TSA background check No medical certification, but must not have a physical of mental condition that would interfere with the safe operation of the vehicle Provision for existing pilots to obtain immediate temporary certification for August 29th

18 New Regulations as of August 29th Operation in Daylight hours, but Civil twilight allowed with proper UAS lighting Visual line of sight operation only Not required to use Visual Observer, but has provisions for doing so. VO may view the aircraft when pilot not able Cannot operate multiple aircraft at once (no swarming) Must operate in VFR conditions (500 ft. below cloud, etc.)

19 New Regulations as of August 29th Provides for authorization from Air Traffic Control to operate in Class B, C, D, E airspace Timeframe and qualifications are unknown, similar to COA? Waivers possible for almost all of Part (a): The Administration may issue a certificate of waiver authorizing a deviation from any regulation specified in If the operation can safely be conducted... List of waivers: Moving vehicle, Night Operations, Visual Line of Sight, Multiple systems, operation over people, ground speed, visibility, altitude, etc.

20 New Regulations as of August 29th Waivers 333 holders for closed set filming will receive waivers for ops over people Similar process as a 333 or COA? FAA plans to post information on waivers granted

21 Other Regulation Updates April 4 th, 2016: Micro UAS Committee 4 categories of drones with some granted permission to fly above people using safety specifications not weight. Advanced to an NPRM July 2016, FAA Reauthorization Bill Passed by Senate with over 30 pages on UAS usage.

Keystone UAS Timeline December of 2014: Began preparing 333 exemption January of 2015: Exemption submitted May 1 st, 2015: Exemption granted Keystone is approved for all types of mapping, remote

22 Keystone UAS Timeline December of 2014: Began preparing 333 exemption January of 2015: Exemption submitted May 1 st, 2015: Exemption granted Keystone is approved for all types of mapping, remote sensing and imagery using Altavian F6500 July 2015: Keystone files for amendment for Quadcopter UAS exemption Oct. 2015: Amendment approved to fly Quadcopter Oct. 2015: Keystone submits a second amendment for more aircraft types Oct. 2015: Keystone submits two civil COAs April 2016: Keystone submits civil COA for bridge inspection June 2016: Keystone receives third amendment approval for all UAS types June 2016: COAs submitted in October approved June 2016: New Part 107 released with implementation in September

Risk Mitigation Keystone uses experienced pilots incurring costs of training flights and manufacturer training Spent extensive time developing its Flight Operations Manual and Standard Operating

23 Risk Mitigation Keystone uses experienced pilots incurring costs of training flights and manufacturer training Spent extensive time developing its Flight Operations Manual and Standard Operating Procedures for UAS operations Checklists for deployment, pre-flight and postflight are always used and updated Weather, wind speeds, etc. are monitored and a UAS risk assessment worksheet is filled out prior to a mission

24 Advantages\Disadvantages

$Life\Range Maximum altitude of 400 feet reduces scale options Payload capability <55lbs Most sensors are not 4 Band Ground control is still necessary for$

25 Disadvantages Drones are NOT a tool for State, County, or even City mapping projects Short comings: Beyond Line of Sight not legalized Short Battery Life\Range Maximum altitude of 400 feet reduces scale options Payload capability <55lbs Most sensors are not 4 Band Ground control is still necessary for most UAS

26 Advantages 1) Resolution: Flying at 1200ft with a Cessna and large format camera 2cm GSD is possible Most UAS data will be acquired at <2cm GSD Rule applies for density of LiDAR as well Quality will vary based on sensor often quality needs to be made up through quantity 2) Frequency\timeliness Able to fly a site repeatedly, on specific dates under most cloud conditions and under/beside for unique views 3) Costs are generally less, but for larger areas the prices are similar and for remote areas it can be much more 4) UAS can be ideal for small scale\high resolution 3D solutions

27 Current Products

$available: precision for mapping\engineering quick and$

28 Current Products Ortho Photo (basemap) with Two types available: precision for mapping\engineering quick and dirty

31 Current Products 3D point clouds: From photo and LiDAR

32 Current Products Bare Earth DEM with raw Contour Lines

33 Current Products Hillshade Flow Accumulation Modeling

$Draping\merging of point$ cloud and ortho photo to

34 Current Products 3D Meshes Draping\merging of point cloud and ortho photo to create a 3D scene Often referred to as 2.5D

35 Current Products True 3D: Use of vertical and oblique imagery to create a true 3D environment

37 Current Products Thermal: Limited to daytime flight Current applications are limited to power line inspection, crack inspection and indoor use. FLIR partnering with DJI is a major step

Current Products LiDAR: Great for determining height values and

derived clouds Accuracy uncertain at this point and vary greatly by

for a $200,000 system to be mounted on a UAV?

38 Current Products LiDAR: Great for determining height values and generating point clouds nearly as dense as photogrammetrically derived clouds Accuracy uncertain at this point and vary greatly by device Cost for highest accuracy systems is prohibitive Insurance for a $200,000 system to be mounted on a UAV? Systems: Velodyne VLP-16 Puck ($8,000) Reigl VUX-1UAV ($200,000) plus RiCopter UAS YellowScan LiDAR

39 Current Applications Package Delivery?

Current Applications Precision Agriculture Uses RGB, multispectral imaging to determine crop health, pest infestation and water runoff/availability Normalized Difference Vegetation Index (NDVI)

40 Current Applications Precision Agriculture Uses RGB, multispectral imaging to determine crop health, pest infestation and water runoff/availability Normalized Difference Vegetation Index (NDVI) derived from near Infrared imagery to assess plat growth/health Timeliness and repeatability are the attractiveness Market is flooded with systems and solutions, but long term usage is unclear will farmers fly the systems themselves? Who will analyze the data? Will there be enough ROI to justify costs?

41 Current Applications Inspection for Bridges, Piers, Towers, Windmills, etc. Keystone has partnered with WSP\PB to provide this service to DRBA Provides access to areas not previously accessible Saves time and money when performing exterior/interior crack inspection Lower cost may allow for more frequent inspections (but will not replace manual inspections completely) Able to record what was seen for reference later (documentation/studies) Reduces risk to inspection crews Soon will be able to map the surface of the road and towers above traffic

43 Current Applications Stockpiles: Landfills, Mines, Quarries for Volumetric Calculations Safer than sending surveyors into dangerous location Cheaper than manned flights Repeatable under most weather conditions Accurate enough for inventories

Current Applications Construction Job Sites: Planning stage: Typical contours, orthos, height models and mapping During: Monitoring of progress for job site updates and client

44 Current Applications Construction Job Sites: Planning stage: Typical contours, orthos, height models and mapping During: Monitoring of progress for job site updates and client updates For As Built documentation of utilities and underground services After completion: Complements 3D laser scanning for BIM (15min in air as opposed to days on the ground)

45 Current Applications Automated 3D Building Extraction Can be done from LiDAR or imagery point clouds Automated process used for building blocks for true 3D Can be joined with parcel data, building attributes, etc. in a 3D GIS

of the events around the visit This application of UAS data is only limited by regulations and will explode due to

46 Current Applications 3D GIS Bentley Systems: Software to create 3D reality mesh Worked with helicopter flyer to capture and generate a 3D model of Philadelphia for the Papal Visit ESM Productions used the model in their planning of the events around the visit This application of UAS data is only limited by regulations and will explode due to the cost saving vs. helicopter for small areas Read more:

Current Applications Advantages of using 3D in GIS: Better predictive flood mapping Which floors will be flooded Viewshed analysis for aesthetics, line of

47 Current Applications Advantages of using 3D in GIS: Better predictive flood mapping Which floors will be flooded Viewshed analysis for aesthetics, line of sight and energy consumption shadows, airflow, reflectance Treat Dome for military and other threat analysis We live in a 3D world, our GIS should be in 3D as well!

$Current Applications 3D GIS ArcGIS\Drone2Map 3D Analyst for ArcGIS delivering some of the same tools as Skyline, but user must provide 3D data.$

48 Current Applications 3D GIS ArcGIS\Drone2Map 3D Analyst for ArcGIS delivering some of the same tools as Skyline, but user must provide 3D data. Use of UAS could be cost effective to generate 3D for small areas Drone2Map (partnership with Pix4D) application from ESRI to turn vertical and oblique UAS imagery into orthos, point clouds and meshes for ingestion into ArcGIS D2M: Still producing 2.5D, but moving toward 3D and generating an appetite for planning in 3D

49 Current Applications Mapping: Currently limited by regulations and tools available For small applications, however, UAS data offers new levels of detail

50 Mapping Accuracy Testing

Keystone Accuracy Testing Keystone flew a privately owned barn in northeastern, PA in December of 2015 with two UAS. Each flight lasted approximately 15 minutes.

51 Keystone Accuracy Testing Keystone flew a privately owned barn in northeastern, PA in December of 2015 with two UAS. Each flight lasted approximately 15 minutes. Flight was in a circular pattern with alternating elevations. DJI X5 camera was approximately 35 ft. above terrain Sony A7R was approximately 68ft. above terrain Simultaneously, 38 control points were surveyed on the barn and ground

Equipment Used DJI Inspire 1 Pro with a DJI Zenmuse X5 camera 16 MP with a 15 mm lens Unobstructed views SteadiDrone Mavrik with a Sony A7R camera 36 MP with a 35mm lens PixHawk Autopilot

52 Equipment Used DJI Inspire 1 Pro with a DJI Zenmuse X5 camera 16 MP with a 15 mm lens Unobstructed views SteadiDrone Mavrik with a Sony A7R camera 36 MP with a 35mm lens PixHawk Autopilot customizable platform Pixel size for DJI averaged 0.10 inch NADIR and averaged 0.17 inch on barn Pixel size for the Sony averaged 0.10 inch as well but was 0.22 inch on barn as it averaged 137 ft. from barn

Datugram3D Workflow Image Loading The Datugram3D software has a simplified workflow of loading images, georeferencing images, measuring/data collection, approving the collected features and exporting

53 Datugram3D Workflow Image Loading The Datugram3D software has a simplified workflow of loading images, georeferencing images, measuring/data collection, approving the collected features and exporting the measurements. Before ingesting the images into the software, redundant images, those with blur and inconsistent lighting were removed from each set. After this process 31 images were ingested from the X5 set and 30 from the A7R set.

Datugram3D Workflow Georeferencing Control points are measured in several images, preferred to be on images at 90 degree angles Solution for entire dataset is generated using tie points and control

54 Datugram3D Workflow Georeferencing Control points are measured in several images, preferred to be on images at 90 degree angles Solution for entire dataset is generated using tie points and control points Image reference quality is indicated by an adjustable color system to show levels of matching quality. User defined levels set by preferences. The default value of 3 pixels for green was used in this case, with most points generating less than 2 pixels of latency.

Feature is projected on all images If the automated feature detection fails,

55 Datugram3D Workflow Measurement Points, lines or polygons are drawn using a single image Feature is automatically detected on at least two other images. Feature is projected on all images If the automated feature detection fails, the user manually places the feature at the appropriate location in other images.

Datugram3D Workflow Approval In the approval stage, each feature can be confirmed in multiple images and more measurements can be made from suggested

56 Datugram3D Workflow Approval In the approval stage, each feature can be confirmed in multiple images and more measurements can be made from suggested locations. Error estimation of object accuracy is available to view in coordinate system units and saved to a file before continuing to the data export stage.

57 Testing - Methodology Test 1 for both X5 and A7R: Test 2 for both X5 and A7R: Georeference Stage Georeference Stage For X5 and A7R 15 points on barn and 2 For X5 and A7R 7 points on barn and 2 ground ground points points Each point was measured in at least 2 images Each point was measured in at least 2 images Same points used for each camera type Same points used for each camera type Measurement Stage Measurement Stage For X5 and A7R 19 points on barn For X5 and A7R 19 points on barn measured as measured as features features Points were refined in Approval stage Points were refined in Approval stage Ability to refine points in this stage improved results

58 Results Control Points The control point estimations and measured points were exported and then compared to the surveyed locations. The points used as control had approximately 0.5 to 1 pixel Absolute Mean Error Control Point Residuals (Inches) Absolute Mean Min Max Camera Test Type X Y Z X Y Z X Y Z DJI X5 High Control DJI X5 Low Control Sony A7R High Control Sony A7R Low Control Mean, Min, Max of Absolute Residuals in Inches (Comparison of Surveyed Location and Measured) for Control Points

59 Results Check Points All of the tests resulted in absolute error values of 0.60 inch or less in X, Y and Z. The Sony A7R tests generating 1 or 2 pixel accuracy for both types of tests. With residuals evenly distributed among X, Y and Z, it is a stable and impressive solution. Note: DJI X5 pixel size of 0.17 inch and Sony A7R pixel size of 0.22 inch DJI MP of 16 is not recommended by Datumate, but still tested well With limited time of trial, blunders were not investigated. Check Point Residuals (Inches) Absolute Mean Min Max Camera Test X Y Z X Y Z X Y Z DJI X5 High Control DJI X5 Low Control Sony A7R High Control Sony A7R Low Control Mean, Min, Max of Absolute Residuals in Inches (Comparison of Surveyed Location and Measured) for Check Points

60 Results

61 Results

62 Conclusions Despite some portions of the project being less than ideal: ambient lighting changes, control placement, etc., the Datagram3D software was able to generate highly accurate data. For the Sony it is evident that an adequate number of control is necessary for best results, but an excessive amount results in rigidity of the solution and poorer results. This is likely due to the changing geometry of the small format cameras requiring flexibility in the bundled solution (especially in software that is not performing self calibration or using additional parameters) when high resolutions are possible. By contrast, the lower resolution (and possibly less stable geometry) from the DJI X5 needs larger amounts of control to limit mismatches and increase overall accuracy within the solution. Know Your Camera, Know Your Software and Test!

63 Future Developments

Near Future LiDAR higher accuracy/lighter weight Miniaturization of Inertial Measurement Units and GPS for removing the need for Ground control Longer

64 Near Future LiDAR higher accuracy/lighter weight Miniaturization of Inertial Measurement Units and GPS for removing the need for Ground control Longer battery life and heavier payloads Detection using other sensors (pipeline leak detection for example) Multiple sensor payloads Near real-time mapping Drone Deploy

Not Too Distant Future Sub-millimeter pixel sizes and accuracy Beyond Line of Site operations will be commonplace: Amazon and Google will spearhead this but the mapping community will benefit Large

65 Not Too Distant Future Sub-millimeter pixel sizes and accuracy Beyond Line of Site operations will be commonplace: Amazon and Google will spearhead this but the mapping community will benefit Large aircraft for mapping Commercial version of Predator (I would call it Dove ) Persistence Solar power, tethered, mid-air refueling, auto battery exchange, etc. Ubiquitous News media will make a drone above you a chance for your 15 minutes of fame, not an invasion of privacy Swarming multiple coordinated UAS

66 Thank You! Contact Us David Day:

LAUNCHING YOUR UNMANNED AIRCRAFT PROGRAM

LAUNCHING YOUR UNMANNED AIRCRAFT PROGRAM THE DARTDRONES TEAM UNMANNED AIRCRAFT APPLICATIONS AERIAL INSPECTIONS AERIAL INSPECTIONS Safer and faster alternative to visual inspection by an individual Damage