Application of Coefficient of Variance for Population Stand Monitoring of Cane Using

Canopy Reflectance-Based Vegetation Index Readings

Saoli Chanda1, Jeffrey Hoy2, Payton Dupree1, Brandon White1, Murilo de S. Martins1, Marilyn Sebial Dalen1, Tapasya Babu1,

Daniel Ernesto Forestieri Muñoz3, and Brenda Tubana1

(1)School of Plant, Environmental, and Soil Sciences, LSU AgCenter, Baton Rouge, LA,

(2)Department of Plant Pathology and Crop Physiology, LSU AgCenter, Baton Rouge, LA,

(3) Escuela Agrícola Panamericana El Zamorano, Tegucialpa, Honduras

Remote sensing is a useful technique for measuring crop

production parameters faster than conventional method.

It uses canopy reflection property of plants to calculate

vegetation indices (VI) that can be used for predicting crop

yield.

The use and application of remote sensing technology in

energy cane (Saccharum sp.) production is understudied.

There have been interest in introducing remote sensing

technology to improve nitrogen fertilizer management and

harvesting logistics.

Planting material may effect the population, millable stalk

and ultimately yield of the cane.

Coefficient of variance (CV) is the statistical measurement of

dispersion and can be calculated as :

Coefficient of variance among VI collected from whole stalk

(WS) or billet (Bi) planted crops can give an estimation of

variability of crop stand.

To establish the relationship between early plant stand

population, tiller number and number of millable stalks.

To determine if CV among plant canopy reflectance based

vegetation index (VI) readings is related to plant stand and

number of millable stalks.

INTRODUCTION

OBJECTIVES

RESULTS AND DISCUSSION

Type of soil: Commerce silt loam (fine-silty, mixed,

nonacid, thermic Aeric Fluvaquent)

Experimental design: Split plot in randomized complete

block design with four replications

Main plot: Two planting materials (billets and whole stalks)

Subplot: six varieties (Ho 02-113, US 72-114, Ho 06-9001,

Ho 06-9002, L 01-299, L 03-371)

Planting method: whole stalk planting was done by hand

and billets were planted mechanically

Fertilizer application: N, P2O5, K2O: 112-0-78 kg ha-1

Planting date: Sept. 2012 (site A) and Oct. 2013 (site B)

Data collection: Canopy reflectance reading and

population count were done 1 (site A) and 5 months (site

B) after planting (once in a month) until tillering. Whereas,

tiller number was determined from May to June at one

week interval up to maximum tillering stage.

Canopy reflectance reading was collected with 2- and 4-

band GreenSeekers® handheld sensors.

Stalk count and measurement of parameters for canopy

characterization were done in the month of July.

Vegetation indices used to calculate CV were:

NDVIred :𝝆𝑵𝑰𝑹−𝝆𝒓𝒆𝒅

𝝆𝑵𝑰𝑹+𝝆𝒓𝒆𝒅 NDVI710 :

𝝆𝑵𝑰𝑹−𝝆𝟕𝟏𝟎

𝝆𝑵𝑰𝑹+𝝆𝟕𝟏𝟎

NDVI735 :𝝆𝑵𝑰𝑹−𝝆𝟕𝟑𝟓

𝝆𝑵𝑰𝑹+𝝆𝟕𝟑𝟓

Billets-planted cane Whole stalk-planted cane

CV

of

ND

VI r

ed

CV

of

ND

VI 7

10

CV

of

ND

VI 7

35

‘1000 s

talk

pop

ula

tion

ha

-1

Parameters Site A 2013 plant cane Site B 2014 plant cane Site A 2014 1st ratoon Billets-planted Whole stalk-

planted

Billets-planted Whole stalk-

planted

Billets-planted Whole stalk-

planted

Plant population ‘000 ha-1 91 60 48 51 - -

Plant population ‘000 ha-1 411 398 286 308 366 331

Stalk population‘000 ha-1 219 181 158 169 151 158

Table 2: Relation between plant population, tiller number, and planting methods at different growth stages of cane.

Both sugar cane varieties (L 01-299, L 03-371) had wider

leaf blade and stalk diameter while all energy cane varieties

were taller. Lower leaf angle and length-where-leaf-bends

values are expected in droopy canopy structure.

During early population count, the plants were too small to

produce any significant relation with CV among NDVI.

Early tiller population showed highest correlation with CV

among different NDVI. As the population increased, the

variability between the crop stand and CV values decreased.

As a result, sensitivity of the relation also decreases.

Coefficient of variance has negative correlation with plant

stand. Over all, CV of NDVIred provided better correlation

with tiller number in both whole stalk- and billet- planted

cane.

Tiller population of both whole stalk- and billet- planted

cane have positive and similar correlation with millable

stalk population.

In site A, billets-planted cane had 52% higher plant

population and 21% higher millable stalks. On the other

hand, both planting materials performed similar in site B

Leaf Angle Measurement

MATERIALS AND METHODS

Canopy Characterization of Different Energy Cane Varieties

Plant Height Measurement Stalk Diameter Measurement

Length Where the Leaf Bent

Measurement

Pla

nt

Hei

gh

t M

easu

rem

ent

leaf Angle

Table 1: Parameters relating to canopy characteristics of different cane variety

Variety Height

(cm)

Leaf Angle

(degree)

Total Leaf Length

(cm)

Leaf Length Bent

(cm)

Leaf Width

(cm)

Stalk Diameter

(cm)

Ho 02-113 143 64 128 114 1.20 1.36

US 72-114 148 69 136 117 1.19 1.44

L 01-299 134 67 129 121 1.65 1.83

L 03-371 119 58 137 91 1.63 2.10

Ho 06-9001 178 70 157 92 1.33 1.26

Ho 06-9002 177 69 238 88 1.44 4.23

Planting of billets Planting of whole stalks Sensing of energy cane

y = 0.0001x2 - 0.128x + 48.761

R² = 0.3656

0

10

20

30

40

50

60

0 200 400 600 800 1000

'000 tiller count ha-1

Plant tiller count and CV of NDVIred

Site A 2014 1st ratoon Site A 2013 plant cane Site B 2014 plant cane

y = 9E-05x2 - 0.0929x + 38.68

R² = 0.1833

0

5

10

15

20

25

30

35

40

45

0 100 200 300 400 500 600 700 800

'000 tiller count ha-1

Plant tiller count and CV of NDVIred

Site A 1st ratoon 2014 Site A 2013 plant cane Site B 2014 plant cane

B

y = 0.0001x2 - 0.1264x + 52.102

R² = 0.323

0

10

20

30

40

50

60

70

0 200 400 600 800 1000

'000 tiller count ha-1

Plant tiller count and CV of NDVI710

Site A 1st ratoon 2014 Site A plant cane 2013 Site B plant cane 2014

y = -0.0275x + 29.97

R² = 0.1694

0

5

10

15

20

25

30

35

40

45

50

0 100 200 300 400 500 600 700 800

'000 tiller count ha-1

Plant tiller count and CV of NDVI710

Site A 1st ratoon 2014 Site A plant cane 2013 Site B plant cane 2014

y = -0.0282x + 37.415

R² = 0.2381

0

5

10

15

20

25

30

35

40

45

50

0 200 400 600 800 1000

'000 tiller count ha-1

Plant tiller count and CV of NDVI735

Site A 2014 1st ratoon Site A plant cane 2013 Site B plant cane 2014

y = -0.0145x + 31.248

R² = 0.0661

0

5

10

15

20

25

30

35

40

45

0 100 200 300 400 500 600 700 800

'000 tiller count ha-1

Plant tiller count and CV of NDVI735

Site A 2014 1st ratoon Site A plant cane 2013 Site B plant cane 2014

F

y = 0.2507x + 85.97

R² = 0.3301

0

50

100

150

200

250

300

350

0 100 200 300 400 500 600 700 800

'000 plant population ha-1

Plant tiller count and millable stalk

population

Site A 2014 1st ratoon Site A plant cane 2013 Site B plant cane 2014

G

y = 0.2608x + 80.299

R² = 0.31

0

50

100

150

200

250

300

350

0 100 200 300 400 500 600 700

'000 plant population ha-1

Plant tiller count and millable stalk

population

Site A 2014 1st ratoon Site A plant cane 2013 Site B plant cane 2014

H

2-band GreenSeeker® Wavebands: red = 660 nm

near infrared = 770 nm

4-band GreenSeeker® Wavebands: red = 660 nm

near infrared = 770 nm

red-edge = 710 and 735 nm

Figures (A-B): Relation between tiller number and CV of NDVIred in billets (A) and whole stalks- (B) planted cane;

Relation between tiller number and CV of NDVI710 in billets- (C) and whole stalks- (D) planted cane; Relation between

tiller number and CV of NDVI735 in billets- (E) and whole stalk- (F) planted cane. ; Relation between tiller number and

stalk count in billets- (G) and whole stalk- (H) planted cane.

A

E

D C