Vegetation Analysis Sensors

Why do we use specialized vegetation analysis sensors?

Visible light aerial photography can often be of some value in vegetation analysis by providing an aerial perspective in a format that can be interpreted intuitively. Areas of heavy weed infestation, severely damaged vegetation, and some types of infrastructure can often be detected and evaluated from visible light images, as demonstrated in this example:

vis-corn-pivot

 

The value of visible light images is, however, limited when it comes to quantitative vegetation analysis. Sensors designed for quantitative analysis of vegetation are usually based on the differences of light absorption by plant pigments within the visible (VIS) to near infrared (NIR) wavelengths. The most prominent light absorption difference in plants, and the most often used for analysis, is based on the relatively strong reflection of NIR (>700 nm) light compared to a relatively strong absorption of VIS (400-700 nm) light. The difference between reflectances measured at these two wavelength ranges provides an indication of the photosynthetic potential of plants. The reflectance difference is typically normalized by dividing it with the sum of the two, resulting in the well known vegetation index known as the NDVI (Normalized Difference Vegetation Index):

 

NDVI = (NIR-VIS)/(NIR+VIS)

 

NDVI is a useful tool for the detection and quantification of vegetation differences, particularly when applied to images that have been combined into an orthomosaic that covers a whole crop production unit, such as a field or irrigation system. Following the calculation of a vegetation index, the results are usually displayed according to a color scale where different colors represent different values to make visual interpretation easier. The underlying data also provide opportunities for quantitative analysis, as demonstrated in the example below:

ndvi_example
quantitative-example

 

Many specific factors may influence photosynthetic potential, and therefore the NDVI, but they can be broadly classified as factors affecting biomass and/or growth vigor. Since biomass and growth vigor are closely tied to productivity in most crops, the NDVI provides a means for assessing crop performance.

 

An ideal vegetation analysis sensor provides a high degree of differentiation between VIS and NIR (ie provides pure bands), a high degree of radiometric dynamic range (how wide the differences are that can be quantified), a high degree of radiometric resolution (how small the differences are that can be quantified, and a high degree of spatial resolution (how close can two objects be and still be separated in the image). In addition, particularly if the sensor will be used on small unmanned aircraft systems (sUAS or drones), the sensor should be small, light weight, robust, and affordable enough for use in small aircraft. There are a large and increasing number of vegetation sensors on the market, and they vary in usefulness and quality, but there is no such thing as an ideal sensor. The design of any vegetation analysis sensor involves a compromise between the factors mentioned above, and the best sensor for a project will be the one that strikes the most useful balance between factors, and compromises less on those factors that are most important for a particular project. In other words, there is no such thing as a “best” vegetation analysis sensor. The optimal sensor choice depends on the needs of the project. That said, some vegetation analysis sensors generally perform better than others in a wide range of applications because they manage to excel in important areas without compromising too much in any area, and choosing the sensor wisely can often be a critical factor in the success or failure of a project, or of a business venture based on providing vegetation analysis services.

A good vegetation analysis sensor option

Arrow Consulting LLC has been involved in quantitative vegetation analysis for research and commercial applications for over 8 years. Lack of good and affordable sensor options in the marketplace drove us to develop our own custom camera conversions, optimized for use in low altitude remote sensing and quantitative analysis of vegetation.

 

A good general purpose option is the Arrow Consulting LLC-designed, field and research-proven Sony a5100 camera, converted to NIR/G/B. It has superior signal to noise ratios due to its large, APS-C sized sensor paired with custom filters that are optimized for consistent light filtration across the sensor, full manual camera control options, and the ability to shoot in RAW and/or JPEG. High resolution (24MP), without compromising image quality, provides freedom for image resolution adjustments to suit a variety of projects.

Sony a5100 kit

Key features that make it an optimal choice for use on small unmanned aircraft:

  • Filters are custom-made and professionally installed to provide high-purity visible and NIR bands, resulting in excellent signal to noise ratios when calculating vegetation indices. Band selection is consistent across the whole sensor, even when used with wide angle (16 mm or 20 mm) lenses. 
  • Superior image quality and dynamic range compared to point and shoot cameras, yet relatively small and light weight compared to DSLR cameras.
  • Bright LED screen for easy adjustment and verification of settings in the field, without the complexity and frustrations associated with WiFi connections.
  • The LED screen magnet is removed to reduce interference with autopilot magnetometers.
  • Includes a Sony Multiport cable and Pixhawk-compatible trigger from Mobilexcoter for reliable triggering during mapping flights.
  • A prime lens provides reliable performance under field conditions.

A custom ready-to-fly converted Sony a5100 with a Pixhawk-compatible camera trigger module, 16 mm prime lens, and magnet removal to reduce interference with autopilot magnetometers, can be supplied for $1,349 plus shipping costs. Please contact us for payment and shipping options.