Greater Control
For Bolder Decisions

Delivering better insight for better grower results

NEC Agricultural Solutions: Agricultural Solutions

Our Technology

Our system takes the data from multiple sensors, and creates various models, each of which describes a physical aspect of your field:

Soil Model Layer

Soil Model Layer
  • We start with a model of the field, including physical parameters such as geographic location, and direction and configuration of the beds
  • We take soil samples at three depths from the field before transplantation, and analyze texture as well as 32 chemical and nutrient properties
  • From this, we create a computer model of the soil in each field
  • At this stage, the model is still quite rough, because it is based on only a few data points

Water Transport Model Layer

Analyze
  • The water transport model predicts how water and fertigation flows from the irrigation system to the plant root system
  • This model layer starts by replicating the design of the physical parameters of the irrigation system
  • In this example, the field has four irrigation blocks which can be independently controlled
  • We install wireless soil sensors in the field, which measure the soil moisture at 5 cm depths to 45 cm
  • The model is initialized by using the soil texture composition from the soil sampling
  • At this stage, the layer, too, is quite rough, because it is based on only a few data points
  • The readings from the soil moisture sensors coupled with the actual irrigation data allows us to refine the model in steps

Weather Model Layer

Decide
  • The weather model layer uses weather data to calculate the evapotranspiration (ET) rate of each section of the field
  • We install a weather station next to the field, and using temperature measurements from the soil sensors, and imaging data, we calculate the ET rate for each block of the field
  • The weather forecast information is used to predict ET rates in the coming days

Imaging Layer

Decide

  • We use successive images to determine the growth rate of each plant and the nitrogen content of the foliage
  • We use satellite images gathered every two weeks, and supplement it with UAV imagery for higher resolutions
  • By analyzing the image square-meter by square-meter, we are able to determine the growth rate compared to the expected growth rate for the planted varietal

Crop Model Layer

Decide
  • Our crop model is based on an open source crop model, with enhancements to utilize our complete set of layers
  • The crop model integrates all of the other layers: soil, water transport, weather, imaging analytics
  • Like the other layers, the model in the early part of the season is rough
  • As we gather data from each set of sensors and inputs, the model becomes more and more precise
  • The crop model is used to predict the harvest yield, by modeling each tomato plant
  • The accuracy of the model and its prediction grows as we gather more data from the field
  • In addition the data from one year to the next is accumulated, making each successive crop modeling in the same field even more accurate

Analytical Decision Support

Imaging Layer

Analytical Decision Support

  • Our analytical engine integrates data from each of these layers of the model and generates
    • Water stress map
    • Nitrogen stress map
    • Forecast yield map

Crop Model Layer

Analytical Decision Support


  • Water stress is defined as the plants requirement for water compared to the availability of water
  • The crop model specifies the needs for water for each plant based on the growth stage of the plant

Analytical Output: Irrigation Recommendation

Irrigation Recommendation
  • It may be relatively standard farming practice to determine the right amount of water for either a consistently high stress block or for a consistently low stress block
  • Since irrigation can be controlled only for an entire block, the difficulty is in determining the proper amount of irrigation not only for the low stress area but also for the high stress area
  • Our system automatically determines the optimal amount of water in each block which will maximize the grower�s objective for the entire block � including high stressed areas and low stress areas

Analytics for Water Transport Modeling

Water Transport
  • Although there are few moisture sensors in the field, we use the crop model and the water stress model to feed back to the water transport model
  • This enables us to have higher accuracy in predicting water levels in each segment of the field