Adatos uses advanced AI and satellite data to provide comprehensive analyses for cotton – including planted area, growth stage, yield forecast, rainfall forecast, moisture content, pest and disease, and nitrogen content.
Growing challenges and opportunities for the cotton industry
Cotton is a major crop with global cotton production projected at 114.4 million bales for 2022/2023. Apart from providing an essential material for fabrics, the cotton industry also provides employment for quarter of a billion people.
Growers must carefully manage cotton plants during the growing and harvest stages – cotton plants require just the right amount of water, nutrients, chemicals and hand-picking at just the right time to ensure optimum quality. The impacts of climate change (including increased water stress, pests and diseases) and the environmental impact of cotton cultivation only add to the challenge of growing this crop. It is all the more important for growers to adopt regular and precise monitoring for resilient and sustainable cotton farming, and for decision-makers to obtain timely, accurate and comprehensive data to develop effective strategies.
With the rise of technologies such as remote sensing and AI, numerous studies have successfully monitored various aspects of cotton remotely. Adatos enables the application of these studies, as well as novel in-house methodologies at low cost, large scale and fine (≤10m) resolution. Below are some analyses that Adatos provides specifically for cotton.
Planted area detection
Adatos can map areas (>0.5ha) planted with cotton over a region of any size. This enables monitoring of the size and distribution of farms, or changes in the agricultural landscape of a large area over time. The AI model is trained with historical cotton plot data, optical satellite data and derived vegetation indices.
Growth stage detection
An overview of the various stages of growth of the crop can aid irrigation, weed control, and pest and disease control planning, as well as harvest date estimations. Growth stages that can be detected include flowering (70-80 days), boll development (100-105 days) and open boll (150-160 days).
Yield forecast
Predicting the cotton yield during first flowering (i.e. about 2 months prior to harvest) allows farmers to make decisions during the growing season. By incorporating the Adatos ML rainfall forecast model, our yield models are able to forecast yield for the upcoming harvest. On its own, the rainfall forecast model also provides valuable insights for planting, irrigation and harvesting plans up to 6 months ahead.
Moisture content
High moisture content during the growing stages is crucial for cotton quality. Precise information of water content within fields enables efficient use of water resources and improved harvest.
Adatos monitors at 10m resolution the moisture content of cotton plants for each month during the growth stage. If sufficient moisture sensor data is available, Adatos can produce absolute moisture content maps and validate the results against the sensor data. In the absence of sensor data, Adatos can monitor relative moisture using optical satellite data and categorize moisture content from low to high categories.
Pest and disease
Cotton farms are affected by a whole spectrum of pests and diseases. Whilst many have been dealt with by the treatment of seeds before planting and genetic engineering, some still require pesticide/fungicide or quarantine, such as the pink bollworm boll weevil and boll rot.
Adatos offers detection and severity assessment of boll rot infected (mid-late stage) cotton plants at high resolution (<2m). If archive imagery is available, historical boll rot data can be used to train the model. Otherwise, new imagery can be ordered and the model trained with new field survey data.
Nitrogen content
Optimum nitrogen levels are crucial for the health and development of cotton plants.
Adatos maps absolute leaf nitrogen content during the growing stage at a resolution of 10m using historical foliar nitrogen data and satellite data. In the absence of nitrogen ground truth data, Adatos can produce relative nitrogen maps based on nitrogen-sensitive vegetation indices.
Have any enquiries?
Adatos builds bespoke models specific to the client’s use case to obtain the most valuable and actionable insights. In addition to the analyses above, Adatos can also develop new models that suit your interests. Send us a message on LinkedIn and explore how Adatos can partner with you.