Multispectral remote sensing provides radically new perspectives on the health and vigor of crops. It allows growers and agronomists to detect areas of stress in a crop and manage these issues immediately. It enables precise application of nutrient inputs and disease preventative actions based on the actual field conditions today. Below are some examples of images and vegetation-index maps using data captured with MicaSense multispectral cameras and analyzed through our data processing tools. Higher vigor areas show as higher numbers. See the FAQ for details on NDVI.
MicaSense ATLAS™ offers a powerful data service for storage, processing, analysis, and presentation of multispectral data. Using secure cloud-based servers, your raw data is processed into actionable information that can be accessed from any internet-connected device.
Fly your mission
Securely upload your raw data
Accurate crop-health maps automatically created
View and manage your data from anywhere
Designed for optimal spectral and spatial resolution and featuring a variety of integration options, RedEdge® is the ideal solution for research and remote sensing professionals.
Powerful imaging in a miniature package: Parrot Sequoia™ is our smallest and most cost-effective remote sensing solution, for integration with any multi-rotor or fixed-wing platform.
Easily integrate RedEdge and Sequoia into multiple types of drones with our custom accessories.
What is multispectral imaging?
The colors we see in light are defined by the wavelength of that light. Plants absorb and reflect light differently depending on this wavelength. Plants typically absorb blue light and red light, while reflecting some green light. They also reflect a much larger amount of near-infrared (NIR) light, which is not visible to the human eye but is visible to multispectral cameras like RedEdge and Sequoia. The reflectance curve of a typical plant is shown below. Reflectance is the percent of light that is reflected by the plant.
By measuring the reflectance of a plant at different wavelengths, multispectral imaging enables identification of areas of stress in a crop, and provides a quantitative metric for the vigor of a plant.
How do multispectral cameras work?
Multispectral cameras work by imaging different wavelengths of light. Professional multispectral cameras have multiple imagers, each with a special optical filter that allows only a precise set of light wavelengths to be captured by that imager. The output of the camera is a set of images for that particular wavelength. These sets of images are then stitched together to create geographically accurate mosaics, with multiple layers for each wavelength. Mathematically combining these layers yields vegetation indices. There are many types of vegetation indices that measure different characteristics of a plant.
Some indices are useful for measuring chlorophyll content of plant leaves, which can provide a reliable indicator of nitrogen status and help inform fertilizer management decisions. Other indices can be used to estimate the amount of leaf area per unit ground area (leaf area index) and help identify within-field differences in crop development or vigor. One popular index is the Normalized Differential Vegetation Index (NDVI), created by combining the reflectance from red and NIR light.
How do professional multispectral cameras differ from single-imager multispectral cameras?
A single-imager multispectral camera uses a blocking filter combined with the standard camera's built-in filter to capture information in 3 wavelengths of light. Because these imagers aren't optimized for remote sensing, the built-in filters are wideband and suffer from data contamination from neighboring bands. This figure shows a comparison of narrowband optical filters and the filters used in a typical single-imager camera.
A professional multispectral camera like RedEdge or Sequoia uses narrowband filters with known characteristics combined with factory calibration parameters, enabling accurate measurements of reflectance that a converted camera simply cannot match.