Home Articles Utilization of Landsat-5 (TM) imagery for sugarcane area survey and mapping in...

Utilization of Landsat-5 (TM) imagery for sugarcane area survey and mapping in Thailand

W. Hadsarang1 & S. Sukmuang 1
1The Office of Cane and Sugar Board, Ministry of Industry, Thailand
Tel : (662) 2023291-4 Fax : (662) 2023293
E-mail Address : [email protected]

Keywords: Landsat, GIS, remote sensing, sugarcane

Abstract
A visual interpretation of Landsat-5 (TM) imagery at a scale of 1:500,000 by the Procom-II projection system was initiated for a survey of selected sugarcane plantations in Thailand in the 1997/98 growing season. The study covered 48 provinces with a total area of 955,255.85 hectares (giving the equivalent production of 42.20 million tons of sugarcane). The Landsat-5 (TM) imagery was shown to be a promising tool for surveying and mapping. The area covered by sugarcane was separated from other crops by a false color composite and it was possible to delineate these in the district level at 1:50,000 map scale. There was a significant difference between the on site survey and visual interpretation. The problem concerning this technique is the clearness of the imagery. During the high sugarcane season, the geographic sectors always have high humidity and cloudy skies hindering image capture, therefore interpretation was accomplished later in the year when the atmosphere was clearer.

Introcuction
Geographic information system (GIS) is recognized as an important tool for geographical surveys and has been applied in agronomic surveys, such as those for sugarcane (Jhoty et al., 1994). The technique has been used in a number of countries such as Australia (Lee-Lovick and Kirchner, 1990, 1991, Lee-Lovick et al., 1992), South Africa (Platford, 1990) and Mauritius (Jhoty et al., 1994). Thailand started a remote sensing project after collaborating in a natural resources survey with NASA in September 1971. The necessary project infrastructure was completed by 1981 and its application for agronomic purpose have been conducted for cassava and sugarcane since 1995. Each year in October, the Office of Cane and Sugar Board, Ministry of Industry must predict the official yield and subsequent price of sugarcane, so that quotas and milling allocations can be set before milling begins in November / December. The total amount of production will influence the price of sugarcane. This paper reports on the results of an application of Landsat-5 (TM) imagery to predict the plantation area of sugarcane and discusses the complication of this technology by comparing it with on-site surveys.

Material and Methods
Imagery from landsat-5 (TM) using film with 4th ,5th and 3rd (red, green and blue) false color composite bands in the ratio 1:500,000 was used. Translation was achieved using a single-band compositor projector (Procom II, Gregory Geoscience Ltd., Canada). The interpretation was confirmed by field surveys using a Global Positioning System (GPS). The drafting film was transferred to a computer system by Scanner (Calcomp Ltd., U.S.A). Using Microstation software, Intergraph Raster Binary (IRASB) and Intergraph Raster Continuous (IRASC) data in the form of vectors was obtained. The plantation area of sugarcane was calculated using software Modular GIS Environmental Analysis (MGA) and Modular GIS Environment (MGE) systems. The map of Thailand (1:50,000) for plotting, was supplied by Royal Thai Survey Department, Ministry of Defense. The results were compared with the on-site surveys of the Office of Agricultural Economics, Ministry of Agriculture and Co-operatives.

Results and Discussions
Visual interpretation of the false color composite was capable of differentiating sugarcane plantation area from other crops such as eucalyptus, pineapple and cassava when the 4th, 5th and 3rd band (red, green and blue) were applied (Figure 1). Using this technique the sugarcane growing area was transcribed onto a map of Thailand at a scale of 1:50,000.

Estimates of the plantation area by visual interpretation and on-site surveys (Table 1) were unfortunately not comparable. The plantation area determined by on-site survey is the one adopted by the Office of Agricultural Economics, Ministry of Agriculture and Co-operatives. The on-site survey was mainly achieved by using a database of cane growers’ registration records, together with interviews in some areas. The registration database was established in 1986, with updated in 1988 and 1992 but these updates only include the addition of new registrants. The lack of continuous updating and ongoing confirmation of the validity of records explains the discrepancy between the two methods.

The results have been instrumental in encouraging the Office of Cane and Sugar Board to review the sugarcane growers’ registration. Application of a Global Positioning System (GPS) confirmed the visual interpretation, confirming the observations that the sugarcane plantation area is decreasing.

The most serious problem encountered by visual interpretation was lack of clarity of the imagery. Sugarcane planting is always at the same time of year, resulting in a high density of leaves and foliage, resulting in a high absorption of light during July/August, but in this period, the sky is always cloudy (Figure 2a), hindering transmission of wavelengths between the visible and infrared preventing penetration to the object. As a result, the imagery for this study was taken from December/January when it was less cloudy (Figure 2b).

As a result of this study, the Ministry of Industry has improved the computerized system including software. The application of data in the microwave region (RADAR) is also being studied. The announcement of the sugarcane plantation area and production in the near future will consider both visual interpretation and on-site surveys.

References

  • Jhoty, I.; Chung, Maryse, Ah Koon, D.; Deville, J. & Ricaud, C. (1994). Digital terrain modelling for the siting of a centre pivot irrigation system in sugarcane. In L.O. Fresco, L. Stroosnijder, J. Bouma & H. van Keulen (eds). The Future of the Land: Mobilizing and Integrating knowledge for Use Options. John Wiley & Sons Ltd., New York.
  • Lee-Lovick, G. & Kirchner, L. (1990). The application of remotely sensed (Landsat TM) data to monitor the growth and predict yields in sugarcane. Proc. Austr. Soc. Sug. Cane Technol., pp 65-72.
  • Lee-Lovick, G. & Kirchner, L. (1991). Limitation of Landsat TM data in monitoring growth and predicting yields in sugarcane. Proc. Austr. Soc. Sug. Cane Technol., pp 124-130.
  • Lee-Lovick, G.; Saunders, M.; Willcox, T. & Bent, M. (1992). The establishment of a geographic information system in the Bundaberg district. Proc. Austr. Soc. Sug. Cane Technol., pp 107-115.
  • Platford, G.G. (1990). A geographic system for use in the sugarcane industry. Proc. S. Afr. Sug. Technol. Ass., pp 83-87.

Table 1. Comparison of estimated sugarcane planted area between the imagery interpretation and on-site survey in 1997/98 growing season.

Provinces Sugar cane area
from the Imagery
interpretation
(hectares)
Sugar area from
the on-site
survey
(hectares)
Discrepancy
(%)
Northern region      
Lampang 5,958.00 6,144.96 -3.14
Chiang mai 174.40 617.92 -254.31
Chiang rai 562.45 140.64 75.00
Uttaradit 12,716.96 15,244.16 -19.87
Sukhothai 36,246.00 30,178.56 16.74
Phrae 1,336.12 4,381.60 -227.93
Phitsanulok 26,910.00 10,139.20 62.32
kamphaeng phet 69,559.00 73,130.56 -5.13
Tak 3,210.00 1,791.20 44.20
Phichit 6,262.00 5,626.56 10.15
Nakhon sawan 59,005.00 62,540.48 -5.99
Phetchabun 20,179.00 6,798.24 66.31
Central plain region      
Kanchanaburi 87,101.00 120,337.12 -38.16
Ratchaburi 27,953.00 51,539.04 -84.38
Nakhon Pathom 9,456.80 15,784.00 -66.91
Suphan Buri 62,499.00 70,429.12 -12.69
Sing Buri 4,428.32 4,792.00 -8.21
Ang Thong 4,083.00 3,886.88 4.80
Chai Nat 7,757.00 6,002.08 22.62
Lop Buri 34,769.76 33,219.68 4.46
Saraburi 9,788.00 6,173.28 36.93
Uthai Thani 10,163.00 18,572.48 -82.75
Prachuap Khiri Khan 7,409.40 12,331.36 -66.43
Phetchaburi 3,778.62 5,948.32 -57.42
Eastern region      
Chon Buri 17,350.00 34,795.20 -100.55
Chachoengsao 3,485.00 13,140.16 -277.05
Prachin Buri 2,001.00 1,189.60 40.55
Rayong 3,201.00 7,410.08 -131.49
Chanthaburi 3,323.00 4,059.52 -22.16
Sakaeo 19,366.00 9,316.80 51.89
North-Eastern region      
Udon Thani 85,960.11 81,346.88 5.37
Nongkhai 4,266.00 1,004.16 76.46
Sakon Nakhon 5,912.00 3,087.52 47.78
Nakhon Phanom 531.60 2,489.60 -368.32
Khon Kaen 66,847.88 43,420.48 35.05
Loei 15,203.00 16,804.64 -10.54
Chaiyaphum 69,182.80 65,947.36 4.68
Maha Sarakham 9,608.00 2,585.76 73.09
Kalasin 27,626.00 10,759.84 61.05
Roiet 4,863.09 1,475.52 69.66
Yasothon 3,669.60 783.84 78.64
Buriram 17,122.92 11,882.72 30.60
Nakhon Ratchasima 69660 65,948.80 5.33
Mukdahan 8,139.00 11,771.68 -44.63
Ubon Ratchathani 120.90
Amnat Charoen 188.00
Nong Bua Lam Phu 5,951.00 4,943.52 16.93
Surin 372.00 258.40 30.54
TOTAL 955,255.73 960,171.52 -0.51

A: SUGARCANE B: EUCALYPTUS
C: FOREST D: PINEAPPLE
E: BARE SOIL F: CASSAVA
G: PRECIOUS STONE MINING

Figure1. False color composite and its interpretation

(a) cloudy

(b) less cloudy

Figure2. Imagery From LANDSAT-5 (TM) at different cloudy levels