Scaling Up Drone Science for Agriculture & Nature Resources through Cooperative Extension

Size: px

Start display at page:

Download "Scaling Up Drone Science for Agriculture & Nature Resources through Cooperative Extension"

Roberta McDowell
5 years ago
Views:

1 Scaling Up Drone Science for Agriculture & Nature Resources through Cooperative Extension Andy Lyons, Maggi Kelly, Sean Hogan, Shane Feirer, Robert Johnson CalGIS 2017, Oakland, CA. May 23, 2017

2 How and why is drone data relevant for agriculture and natural resource management? How does a big organization like Cooperative Extension develop and support drone mapping?

3 Background Under which US President were Land Grant Universities and Colleges first established? A. Lincoln C. FDR B. Teddy Roosevelt D. Nixon

Morrill Act of 1862 transferred Federal land to the States to teach such branches of learning as are related to agriculture and the mechanic arts.

4 Morrill Act of 1862 transferred Federal land to the States to teach such branches of learning as are related to agriculture and the mechanic arts. Hatch Act of 1887 provided funding for Agricultural Experiment Stations. Morrill Act of 1890 required each state to show that race was not an admissions criterion, or designate a separate land-grant institution for persons of color Smith-Lever Act enabled the creation of Cooperative Extension.

5 In 1913, California opened its first county Cooperative Extension office in Humboldt County.

6 US Land Grant Colleges and Universities

7 A bridge between UC scientists and the public Applied research and extension The 11 th Campus of UC 115 campus-based Cooperative Extension specialists 165 locally based CE Advisors and Specialists 57 county offices 9 Research & Extension Centers 8 Statewide Programs 2 Institutes

8 HEALTHY FOOD SYSTEMS HEALTHY ENVIRONMENTS HEALTHY COMMUNITIES HEALTHY CALIFORNIANS

9 Informatics and GIS Program IGIS envisions an ANR where all local and statewide research and extension programs integrate cutting edge geospatial tools into project planning, analysis, implementation, and communication.

10 Team Shane Feirer Lead GIS Robert Johnson GIS Progammer/Analyst Sean Hogan Training, Drones, Flux Maggi Kelly Director Andy Lyons Program Coordinator

(right) collected at the UC Desert Research and Extension Center.

11 Why Drones? High spatial resolution Orthomosaic image (left) and modified normalized difference green index (right) collected at the UC Desert Research and Extension Center. This image was acquired from an altitude of 400 feet above ground level, using a 12 megapixel GoPro camera. The ground sample distance of each pixel is approximately 3.4 inches, over approximately 218 acres.

12 Why Drones? On demand sampling Robert Starnes from UC Davis flies a drone over a field of strawberries in San Luis Obispo County to study how reflectance data may help detect outbreaks of spider mite.

13 Why Drones? Sensors, sensors, sensors LiDAR Multispectral Thermal green red red edge near IR

14 Why Drones? 3D Structure Photogrammetric image stitching Giant sequoia 3D maps produced by multispectral images analyzed with Pix4D software for a study about tree physiology, forest microclimates and carbon stocks.

15 Why Drones? Public Engagement Sean Hogan explains drone science as part of the Farm Smart public outreach program at Desert Research and Extension Center, Imperial County, CA.

16 Challenges for Scaling-Up: The Science Matching data to the question Photogrammetry Data analysis & modeling Remote Sensing

illumination sensor IMU Tablet apps mission planning drone calibration sensor calibration

17 Challenges for Scaling-Up: Technology Drone battery GPS IMU propellers antennae firmware gimbal or USB Wifi or Bluetooth WiFi Sensor gimbal memory card lens sometimes: battery illumination sensor IMU Tablet apps mission planning drone calibration sensor calibration internet connection OS WiFi or USB Flight controller battery antennae manual flight controls

18 Challenges for Scaling-Up: Flight Skills mission planning risk management regulatory compliance equipment checks calibration working with apps launch sequence manual flight skills in-flight trouble-shooting data management

19 Challenges for Scaling-Up: Institutional regulatory compliance safety liability insurance accidents staff development equipment obsolescence

20 Scaling-Up Strategy outreach training mapping services R&D

21 Outreach

22 Training Workshops Intro to GIS - Crop Ag - Forestry - Rangeland ecology Web GIS & Story Maps Remote sensing Mobile Phone GPS Mobile Data Collection Data Management Drone technology & data analysis

23 Training

UC Unmanned Aircraft System Safety System-wide expertise, support and training for regulatory compliance, risk management and the safe operation of drones, across the

24 UC Unmanned Aircraft System Safety System-wide expertise, support and training for regulatory compliance, risk management and the safe operation of drones, across the UC system Authorized to grant certain forms of approval for UAS operations in the US. UC Drone Policy in

25 Drone Fleet 3DR Solo DJI Inspire 1 Pro DJI Matrice 100 DJI Phantom 3

26 Sensors RGB Multispectral Thermal RGB Orthomosaic and mndgi

27 R&D DJI Matrice 100 with custom doublecamera mount

28 Project Sampler

29 Digital Elevation Modeling A digital terrain model derived from a UAS image mosaic in the Bishop Ca. area. The image was acquired from an altitude of 100 feet above ground level, and using a GoPro 12-megpixel digital camera. The pixel resolution is approximately 0.8 inches over approximately 12 acres.

30 Desert Tortoise Habitat

31 Watershed Hydrology 3D point cloud and NDVI model from Hopland REC, January 2017

Forestry 3D Modeling and Classifications A colorized point cloud

highlighting forest stress due to drought and pine beetles (right),

of 300 feet above ground level using a GoPro 12-megpixel digital

32 Forestry 3D Modeling and Classifications A colorized point cloud rendered in Pix4D (above), a Random Forest classification highlighting forest stress due to drought and pine beetles (right), and the classification image draped on a digital elevation model (top right). The image data was collected in El Dorado County, from an altitude of 300 feet above ground level using a GoPro 12-megpixel digital camera. The ground sample distance of each pixel is approximately 2.5 inches. Dead Trees Stressed Trees Healthy Trees Soil and Tree Litter Shadows

Rangeland Classifications Lake Berryessa Cow Fecal Pat Grass Tree UAS color image mosaic (left) and a

digital camera. The ground sample distance of each pixel is approximately 0.

33 Rangeland Classifications Lake Berryessa Cow Fecal Pat Grass Tree UAS color image mosaic (left) and a maximum likelihood classification image of land cover (right) from data collected in the Lake Berryessa CA region. The imagery was acquired from an altitude of 100 feet above ground level using a GoPro 12-megpixel digital camera. The ground sample distance of each pixel is approximately 0.8 inches over an 18 acre area. Note Shadows in the peripheral area outside of the box, not directly underneath the drone s flight lines, were frequently misclassified as fecal pats and thus are separated by the box.

34 Thermal Imaging Find the sheep in this image. From Hopland REC.

35 Applications Rangeland Assessment UAS image and vegetation classification. This image was acquired from an altitude of 215 m above ground level, and using a Canon SD megpixel digital camera. The pixel resolution is 6 cm. Image courtesy of the USDA-ARS Jornada Experimental Range

from UAS imagery collected in the Bishop CA area.

36 Rangelands Bishop Ca. Color infrared orthomosaic image (left) and normalized difference vegetation index (right) from UAS imagery collected in the Bishop CA area. This image was acquired from an altitude of 100 feet above ground level, using a 1.2 megapixel Micasense Sequoia camera. The pixel resolution is approximately 1.45 inches, over approximately 12 acres.

37 Visualization

38 National Network Geospatial Science for Cooperative Extension webinars ask an expert map gallery map tales and more!

39 INSPIRATION Because we can t all grow up to be astronauts. Andy Lyons <andlyons@ucanr.edu>

Capture the invisible

Capture the invisible A Capture the invisible The Sequoia multispectral sensor captures both visible and invisible images, providing calibrated data to optimally monitor the health and vigor of your crops.