Field campaign in Brazil

The Gamaya team has operated on the K Farm in Brazil during the 5 days period in June 2015. All the fields contained corn crops from multiple seed providers: Syngenta, Limagrain and Pioneer. Data was collected by hand, drone and manned aircraft. The current article summarizes the results based on a flight covering the entire farm.

Disease Diagnostic

We applied a range of conventional and advanced proprietary analytical methods in order to identify fields which suffered from a number of issues, including nutrient deficiency, presence of weeds and environmental stresses. Subsequently, we were able to map and classify these problems with a relatively high level of accuracy and precision.

The different plots are ranked according to their relative performance and expected yield, and preliminary recommendations are provided with the aim of improving the overall performance. The expected increase in yield as a result of the implementation of the recommendations provided will range between 5% to 30% depending on the current performance of each particular plot. A cumulative increase of 10% for the entire farm is expected, resulting in a total annual return on investment of 300%.

Problem Overview

Below we’ve have highlighted identified major problems that can and should be addressed starting from the next growing season:

Problem: Strong difference between performance of different corn varieties inside one field. For instance, in the specified field Pioneer seeds work much better than Limagrain.

Recommendation: Optimize distribution of different types of seeds according to soil type and composition of nutrients in the soil.

Problem: Low yield efficiency expected in several fields (in some more than 40% losses) due to a number of different problems, such as diseases and weeds.

Recommendation: Constant monitoring of crop development and presence of weeds and diseases will help to reduce negative impact and increase yield efficiency.

Problem: nutrient deficiencies appear to be an issue across the farm, potentially affecting a production yield.

Recommendation: Chemical analysis of soil not necessarily represents nutrients in the crop. Precise fertilisation in accordance with crop needs and soil conditions.

Yield Forecasting

Yield forecasting is a precious piece of information needed for farm managing. Precise and accurate yield forecast is generally challenging to obtain and has to be derived from many source of information such as soil composition, weather history, fertilization history, etc. Nevertheless, several studies have shown that some vegetation indexes are strongly linked to yield forecasting, in particular the Chlorophyll index — red edge, presented below.

Chlorophyll content is an important property that is linked to crop health and crop yield. It requires fine spectral resolution in this range, as provided by our cameras.

Because this index is also linked to crop health, on the surface it gives globally similar information to NDVI. However, such index can be used complementary with the NDVI to estimate yield forecasting. By using both information, it is possible to create the yield forecast.

Corn Varieties

While still in Brazil, we were easily able to detect the different corn varieties automatically, based solely on the data from our hyperspectral imaging camera. Then we compared efficiency of each corn variety and found out strong difference in their performance even across one field. All this highlights an importance of corn varieties and their selection.

The image below reveals variations in the color of a field, which indicate abnormal growth, crop loss or health problems. The issue was thoroughly investigated and results are provided. Information about abnormal growth and health problems can facilitate in-time and qualitative management decisions throughout the life cycle of crop.

Chemical Analysis

Lastly, we were able to produce chemical maps by correlating chemical data from leaf analysis with our spectral data. The resulting chemical maps span the entire farm and offer an almost X-Ray view of nutrient distribution across the different fields. This information can be used to make a precise fertilization in accordance to crop needs.