|| Home || SEARCH|| Contact || Free Research Publications  ||  Eshop || Privacy Statement ||

Download full report (500k - HINT: right click and save the file to your hard disk before opening)

Using Near Infrared technologies 
to enhance precision management of rice cropsmm

by
Alison Bowman, NSW DPI, Wagga Agricultural Institute, Wagga 
Susan Ciavarella, IREC, Yanco Agricultural Institute, Yanco 
Anthony Blakeney, Cereal Solutions, Lane Cove 
Graeme Batten, SeaSpec, Woolgoolga

January 2007

RIRDC Publication No 06/124   RIRDC Project No DAN-222A

Executive Summary
Importance of this research
This report contains the results of an evaluation of sensors which could increase the precision of variable crop fertilization. This technology could lead to higher returns per hectare and per megalitre of water through the achievement of uniformly high yields of high quality rice. The results would have immediate value to rice growers when incorporated into the SunRice Tissue Testing Service.

Background
Since 1996, over 1500 rice crops have been sampled and sent to the SunRice for near infrared (NIR) Tissue Test analysis and fertilizer recommendations each year. This translates into over 40% of rice growers using the SunRice service in any one year. While this is an extremely high adoption rate for a soil or plant analysis service there are weaknesses of the current system. These include the time required to collect and preserve the samples, the large number of rice growers who do not utilize the service, the minimal sampling intensity relative to the variation across a crop, and the expectation that the fertilizer needs of rice vary widely across a field.

Many rice growers are now using simple Normalised Difference Vegetative Index (NDVI) images of their crops to assist with crop management. These images utilize two wavelength bands – one visible and one infrared. Spectrometers/radiometers which scan many visible (VIS) and NIR wavelengths are available. However, to achieve the maximum benefits from those instruments direct comparisons on sets of crops of known density and nutrient status need to be made.

Objectives/aims of the proposed research
The aims of this project were to provide:

• an evaluation of the potential of available airborne scanning NIR sensors to determine variations in dry matter and shoot composition (eg Nitrogen and starch) across rice crops
• an understanding of the limitations of fresh tissue analysis by airborne NIR sensors
• calibrations for nutrients in fresh rice tissue for use in interpreting data obtained from airborne NIR sensors
• a basis on which to keep the Tissue Testing Service in line with the latest technology.


Methods
Reflectance spectra from a range of rice genotypes, crop ages and densities and from common weed species were collected during the growing season using above crop scanners (hand held, airborne and satellite borne). The spectra were used to develop calibrations for actual shoot biomass, nitrogen (N) and nitrogen uptake (N_uptake) in fresh shoots (as opposed to the current calibrations which use dry and ground shoots). The spectra were also be used to generate a library of signals for actual shoot biomass, nitrogen and other constituents and to identify weeds and unhealthy areas in crops. The expertise of the NIR specialists engaged on this project enabled them to effectively handle the bit-map data generated by red green blue (RGB) and infra-red camera sensors to establish if existing camera data can also be of use for crop nutrient and density prediction.

Key findings
Preliminary calibrations on remotely sensed spectra of rice crops, with correlations of approximately 0.85, were established for % N and N_uptake.

Implications for relevant stakeholders:
The results will be of immediate value to rice growers when the preliminary calibrations produced in this project are improved and extended to include remotely sensed spectra from several more seasons and protocols established to incorporate them into the SunRice Tissue Testing Service.

Recommendations
More remotely sensed images and the related “ground truthing” need to be acquired over a range of seasons, varieties and areas to ensure strong and robust calibrations.

Implications
The future use of remotely sensing the nutrient status of rice crops will depend on the continued availability of suitable sensors, the development of timely protocols for the delivery of data and the continual improvement of calibrations

 


Site map Contact us Feedback Links Disclaimer Privacy Statement

Last updated: January 2007    Copyright RIRDC
http://www.rirdc.gov.au/reports/RIC/06-124sum.html