Agriculture is the lifeblood of the economy in India, the basis on which 70% of the country’s population sustains itself. As such, our country’s economy is largely dependent upon agricultural development. Information related to agricultural production is essential for resource planning and its allocation to different agricultural sectors. Such information can be collected efficiently and reliably using remote sensing technologies, which can map crop acreage, type, and yield. Remote sensing offers structural information with respect to the vegetation’s health and spectral data on the biophysical characteristics of the crop, which assist in the assessment of crop growth and vigor, both of which are directly related to yield.
Multi-date spectral information can provide an accurate picture of the seasonal crop cycle and predict crop yield more accurately, in comparison with the use of single-date spectral data, during the growing season. It is essential to forecast crop yield long before harvest, particularly in areas known to have uncertain climatic conditions. This forecast makes it possible for decision-makers and planners to decide how much to import in the event of a shortfall or, alternately, how much to export should there be a surplus. In many countries, estimation of crop yield is based on traditional techniques of data collection that are based on ground-based field reports and visits. These visits and reports tend to be costly, subjective, time-consuming, and prone to errors resulting from incomplete observations on the ground, which leads to substandard assessment of crop yield and estimation of crop area.
Many studies have been done in India with respect to wheat crop yield forecasting, but very few have been conducted on sugarcane yield prediction, despite this being a cash crop for Northern Indian farmers and the existence of many sugar industries (sugar cane crashers). This article examines a number of spectral indices that could be utilized for crop classification and mapping. It also takes a look at a range of models of crop yield forecasting that use remote sensing, meteorological, and other collateral data, such as Meteorological, Time Series, Spectral, and Integrated Yield Models. A detailed study of literature revealed the gaps in research and the approach that should be adopted for use in sugarcane crop yield forecasting in the Muzaffarnagar and Pantnagar Districts of Northern India, with their respective agro-climatic zones. It is believed that the results of this research will prove extremely useful to India’s food corporations, sugar industries, and Ministry of Welfare.
The primary purpose of this study is the discrimination of sugarcane crop using the temporal single- and multi-sensor data approach as well as the estimation of the yield of sugarcane. Those two objectives can further be broken down into the following sub-objectives:
- Identification of crop spectral growth profile using a temporal and multi-sensor approach
- Consideration of the separability of sugarcane and other crops based on their spectral growth profile
- Estimation of the acreage under sugarcane via satellite images
- Estimation of sugarcane yield and production using VI-LAI based or Agromet-spectral-trend yield models
The proposed methodology is illustrated below:.