uas-platforms for research, phenotyping and precision ... · uas-platforms for research,...

25
UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2), Jinha Jung (3), Anjin Chang (3), Juan Enciso (4), Andrea Maeda (1) (1)Texas A&M AgriLife Research, Corpus Christi, TX, (2)Texas A&M AgriLife Extension , Corpus Christi, TX, (3)Texas A&M University - Corpus Christi, Corpus Christi, TX, (4) Texas A&M AgriLife Research, Weslaco, TX, Precision Cotton Working Group Improving Life through Science and Technology.

Upload: others

Post on 31-May-2020

10 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

UAS-Platforms for Research, Phenotyping and Precision Management

Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2), Jinha Jung (3), Anjin Chang (3), Juan Enciso (4), Andrea Maeda (1)

(1)Texas A&M AgriLife Research, Corpus Christi, TX, (2)Texas A&M AgriLife Extension , Corpus Christi, TX,

(3)Texas A&M University - Corpus Christi, Corpus Christi, TX, (4) Texas A&M AgriLife Research, Weslaco, TX,

Precision Cotton Working Group Improving Life through Science and Technology.

Page 2: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Platforms and Sensors

Data interpretation, Applications

Data analysis & visualization  

Objectives of AgriLife-TAMU-CC Remote Sensing Technology Program

Page 3: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

REMOTE SENSING TECHNOLOGY FOR PHENOTYPING AND PRECISION MANAGEMENT APPLICATIONS

Page 4: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Remote Sensing Equipment: Ground‐based Platform 

Multi Spectrum

Air T & RH SensorsInfrared Canopy Temperature GPS Antenna

Page 5: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

UAS Equipment, Server and Software(2015)

• Structure from Motion (SfM)– Pix4D– Photoscan Pro– Opendronemap

• Geospatial data products– Orthomosaic image– 3D point clouds– DEM

Computational platformIntel i7‐5820K (3.3 GHz) processor (6 cores with 12 threads)64 GB Quad channel DDR4 memoryGTX 960 2GB superclocked GPU with 1,024 CUDA cores

<DJI Phantom 2> Vision Plus

Page 6: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Texas A&M AgriLife Research and Extension CenterCorpus Christi, Texas

Cotton Phenotyping

35 commercial cultivars, replicated 4 times, total of 140 plots.(2 rows, 40’ long), both rows were machine harvested on 08/10

Page 7: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

22 Blocks (1 m2) /Entry(Plant height and canopy cover, NDVI and canopy temperature, open boll count)

Page 8: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

– Mosaic Images– DSM (Digital Surface Model) – DEM (Digital Elevation Model), DSM was generated from 

UAS data– DEM was generated from ground point (about 150 points) selected manually– CHM – Crop Height Map 

<DSM> <DEM> <CHM>

Data Analysis – Plant Height

Page 9: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Fitting sigmoid curves• X‐axis: days after planting • Y‐axis:  Plant Height• Simplest sigmoid equation 

Growth Analysis of Plant Height DevelopmentCultivar 16, Rep. 1.  SeedCotton Yield = 3236 lb/A 

35 InchMax Growth Rate = 4.0 cm

At 57 DAPDuration = 19 Days

(m/day)

Growth Rate (m

/day)

Page 10: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Max Growth Rate = 3.35 cmAt 60 DAP

Duration = 21 Days31 Inch

Fitting sigmoid curves• X‐axis: days after planting • Y‐axis:  Plant Height• Simplest sigmoid equation 

Cultivar 26, Rep. 1.  Harvest SC Yield = 1386 lb./A 

Growth Rate (m

/day)

Page 11: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Spectral AnalysisCanopy Cover 

• Drawing individual crop grid– Yellow line: center line of each row– Red box: individual crop grid 

• Grid sized: 1×1 m

• Binary canopy classification– 3 parameters

• Red/Green, Blue/Green• 2×Green – Red - Green

• Canopy cover estimation- Area of white in each grid /total area

Page 12: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Canopy CoverJune 27, 2015

Page 13: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

0102030405060708090100

0 20 40 60 80 100Days After Planting

Canopy Cover (%)

EXAMPLE OF CANOPY COVER ESTIMATION FROM UAV IMAGESAverage per entry for each date

Page 14: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

1000

1500

2000

2500

3000

3500

60 65 70 75 80 85 90 95 100

Yield 

Canopy Cover (%)

Yield (Lb. SC / Acre) vs Canopy Cover at June 26th

A

B

C

D

F

1000

1500

2000

2500

3000

3500

60 65 70 75 80 85 90 95 100

Yield 

Canopy Cover (%)

Yield (Lb. SC / Acre) vs Canopy Cover at July 2

A

B

C

D

F

Page 15: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

1000

1500

2000

2500

3000

3500

0.76 0.78 0.80 0.82 0.84 0.86 0.88 0.90 0.92

Yield 

NDVI

Yield (Lb. SC / Acre) vs NDVI, June 26th

A

B

C

D

F

1000

1500

2000

2500

3000

3500

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Yield 

CWSI

Yield (Lb. SC / Acre) vs CWSI June 26th

A

B

C

D

F

Page 16: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Monitoring Crop MaturityDow (Phytogen) cultivar evaluation test

• Altitude: 30 m (above ground)• Programmed mission using Pix4D capture app• Processed with Agisoft Photoscan Pro → Orthomosaic &

3D point cloud data• Resulting spatial resolution: 8 mm

07/24/2015 07/29/2015 08/05/2015

Page 17: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),
Page 18: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Low                        HighLint Yield

Page 19: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Plant Height Map, King Ranch, Kingsville, TexasJ. Landivar, S. Searcy and S. Yang, June 13 1997

Precision Management

Precision Pix ApplicationsDefoliation Time and RateNematodesFertility, Etc.

Page 20: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Crop Height Model at Individual Crop Grid Level

<2015/06/03>

<2015/07/02>

<2015/07/29>

<2015/06/11>

<2015/07/10>

<2015/08/05>

<2015/06/18>

<2015/07/17>

<2015/08/14>

<2015/06/23>

<2015/07/24>

Page 21: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Growth Curve Analysis

• Grid level

<Growth Height> <Growth Rate>

Page 22: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Variable Application Rate (ml/ha)

Aggressive Less Aggressive

Page 23: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Conclusions 1. Canopy Cover is an important seedcotton yield component  

2. NDVI and canopy temperature can help fine tune selection of elite genotypes

3. Growth analysis of plant height and canopy cover development appears to be a promising diagnostic tool 

4. Open boll analysis can be an important variable for phenotyping

Page 24: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Enhance Sensors Platform  • Larger Octocopter• Lidar & GPS• Hyper‐Spectral camera• Infra‐red camera 

Data Analysis • Develop software to extract data and analyze seasonal changes in plant height, canopy cover, 

NDVI, and canopy temperature. • Evaluate sensors and develop software to analyze crop maturity, open boll counts and yield

Future Work 2016-17

Data Interpretation• Develop automated procedures to quantify crop responses to experimental treatments and assist 

in the selection elite genotypes   • Develop precision management applications 

Page 25: UAS-Platforms for Research, Phenotyping and Precision ... · UAS-Platforms for Research, Phenotyping and Precision Management Juan Landivar (1), Murilo Maeda (1), Josh McGinty (2),

Thank you

Dr. Jinha JungTAMU‐CC

Dr. Murilo Maeda AgriLife‐Res.

Dr. Josh McGintyAgriLife‐Ext.

Dr. Anjin Chang TAMU‐CC‐AgriLife

Dr. Juan EncisoAgriLife‐Res.

Our Thanks to 

TSSC‐Cotton Incorporated, Dow AgroSciencesShell Oil Company

Texas A&M AgriLife Seed Bioenergy FundsOffice of the Director (AgriLife‐Research)

Ms. Andrea Maeda AgriLife‐Res.