■МНШПИМММММ 
пяашйнммяМННЯЙЯМИСМИЙН! 
In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
In: Wag 
10 
CANASAT PROJECT: MONITORING THE SUGARCANE HARVEST TYPE IN THE STATE 
OF SAO PAULO, BRAZIL 
Daniel Alves de Aguiar 1 , Wagner Fernando da Silva 1 , Bernardo Friedrich Theodor Rudorff 1 , Marcos Adami 1 
'National Institute for Space Research (INPE), Remóte Sensing División (DSR), 
Sao José dos Campos, 12227-010, Sao Paulo State, Brazil. 
{daniel; wagner; bernardo, adami}@dsr.inpe.br 
KEYWORDS: burnt sugarcane; unbumt sugarcane; satellite monitoring; multitemporal analysis; environmental. 
ABSTRACT 
In order to increase the sustainability the ethanol production process from sugarcane, it is necessary to reduce actions that harm the 
environment. One of these actions is the burning of sugarcane straw prior to harvest. In Brazil, agro-environmental protocols between 
the government and the sugar-ethanol sector have been signed. The agro-environmental protocol in the state of Sao Paulo requires the 
total termination of burning by 2017. To meet the objectives of the protocol, remote sensing satellite images are used to monitor and 
inspect the burning reduction each season. Using satellite images, the Canasat project has mapped the harvesting method (with or 
without burning) in the state of Sao Paulo since the 2006/07 season. The objective of this study is to present the methodology utilized 
and to evaluate the evolution of the harvesting method between the 2006/07 and 2008/09 seasons. The harvest type was also 
evaluated for each declivity class. Results demonstrated that the unbumt harvest increased from 50.9% in the 2006/07 season to 
65.8% in the 2008/09 season. In the three seasons analyzed, approximately 97% of the total area available for harvest in the state is 
located in areas with a favorable declivity for harvesting without burning, i.e., mechanically. The western region of the state had the 
greatest expansion in sugarcane cultivated area and also the largest increases in areas of unbumt harvest. 
1 .INTRODUCTION 
Brazil is the world’s largest producer of sugarcane FAO (2009). 
Industrial-scale production occurs primarily in the states located 
in the Northeast and South-Central regions of the country. The 
state of Säo Paulo, in the South-Central region, is the largest 
producer in the country and was responsible for 61% of Brazil’s 
sugarcane production in the 2008/09 season (UNICA, 2009). 
The country is attempting to achieve sustainable ethanol 
production and to obtain its socio-environmental certification 
(Goldemberg, 2007; Goldemberg et al., 2008). One of the 
principle goals in this endeavor is to terminate the burning of 
sugarcane straw prior to harvest. Therefore, in 2007, Sao 
Paulo’s State Secretary for the Environment (SMA) and 
representative in the sugar-energy sector signed the agro- 
environmental protocol for the sugar-ethanol sector. This 
protocol decrees, among other measures, ending the burning of 
sugarcane straw by 2014 in areas that are mechanically 
harvested (declivity less than or equal to 12%) and by 2017 in 
areas that are harvested non-mechanically (declivity greater 
than 12%; http://homologa.ambiente.sp.gov.br/etanolverde 
/english.asp). It is worth to mention that manual harvest of 
sugarcane can only be performed by burning the straw. 
In Brazil, since 2003, the National Institute for Space Research 
(Instituto Nacional de Pesquisas Espaciáis - INPE), the 
Industry Sugarcane Association (UNICA), the Center for 
Advanced Studies on Applied Economics (CEPEA) of the Luiz 
de Queiroz Agricultural School (Esalq/USP) and the Center for 
Sugarcane Technology (CTC) have maintained the Canasat 
project (www.dsr.inpe.br/canasat/eng/). Using remote sensing 
imagery and geoprocessing techniques, the Canasat project 
monitors areas planted with sugarcane. Initially, mapping was 
performed only in the state of Sâo Paulo (Rudorff et al., 2005), 
however since 2005, mapping has been extended to the other 
five states in the South-Central region of Brazil (Rudorff and 
Sugawara, 2007). These six states are responsible for 72.7% of 
Brazil’s sugarcane production. 
One of the project activities, in the state of Sao Paulo, is to 
monitor the type of harvest (with or without burning the 
sugarcane straw) performed since the 2006/07 crop season. 
Information provided by the project has been utilized by both 
the government and private groups. Beginning with the 
2009/2010 crop season, maps depicting the type of harvest are 
generated monthly and sent to the SMA of Sao Paulo State. 
The SMA inspects these maps to determine if the straw burning 
has been authorized. 
The implementation of this protocol has contributed to 
increasing the monitoring and inspection capacity of the sugar- 
energy sector. Using remote sensing satellite images, 
information can be obtained at multiple time-points, and 
therefore providing a monitoring system for the sugarcane 
production process. Furthermore, this crop is generally grown in 
large areas and possesses a long phenological cycle and a long 
harvest period, averaging 12 months and 8 months, respectively. 
These characteristics facilitate crop identification in the images 
(Abdel-Rahman and Ahmed, 2008). 
The objective of this study is to present the methodology of the 
Canasat project, including the monitoring of the harvesting 
method, and to analyze the evolution of the harvest areas with 
and without straw burning from the 2006/07 to the 2008/09 crop 
season. The maps generated by the project may serve as a basis 
for greenhouse gas emission models (Lara et al., 2005), carbon 
storage in silos (Galdos et. al., 2009), public health studies 
(Ribeiro, 2008) and as an aid to public policy in the agricultural 
sector (Moraes, 2007). These maps also allow evaluating the 
area of harvestable sugarcane that was not harvested due to 
weather or industry constrains which is essential information for 
accurate yield estimation. 
Identifie 
burning 
of Säo 
Paulo 
248,20 
and the 
Figure 
well as 
crop ye 
The idt 
visual 
taken 
the T1 
taken 
an alten 
the TM 
of inten 
images 
from TI 
utilizing 
interpol 
Monitor 
map of 
mask fo 
the sug 
crop ye 
beginnir 
Figure 
2.2), the 
declivity