There are three broad types of targets, where machine learning helps to detect things.

While there is no single rule that determines which elements can be detected with what hardware, we find the following three categories to be useful guidelines.


Replaces wet chemistry lab tests

Spectrometers that typically use hundreds or thousands of spectral bands are better at finding invisible elements. If you already have a spectrometer, then that’s the place to start..
Examples: Nitrogen levels in soil, early disease detection, protein or nutrient levels in plants.

(Read more...)

The use of spectrometers to detect everything from nutrients to disease to protein levels to moisture (and so much more), is not new. Search the internet for “NIR spectroscopy for sugarcane” (use your target of choice) and don’t be surprised if a scientific paper has already been published on that exact target. What this means is there is likely a near infrared signal already associated with your target of interest. How you turn that good news into a practical application for your business – is what we are all about.


Replaces or augments expert eyes

Multi-spectral cameras can be used here. Since our analytics engine is great at the combination of spatial data and spectral data, disease or elements that agronomist or pathologists can see can usually be found with a predictive Aapp that uses the training samples identified by the experts in their field.
Examples: Specific disease on a plant leaf, wheat rust, differentiating oats from wild oats.

(Read more...)

The idea of using a camera to replace expert eyes can often increase consistency of results and the ability to work 24/7 can be helpful and the cost savings could be significant. Access to expert eyes can be invaluable to help create properly labeled training data to build the Aapp in the first place.


Replaces manual detection in large areas, like a greenhouse or large field, or repetitive tasks that are tedious, repetitive or time-consuming. Although the target can be seen by eye, it can just as easily be done with a camera and an Aapp. Building a new predictive Aapp is now something you can do yourself – even with a cell phone.

Examples: Invasive species, color tolerances, abnormal shapes or sizes.

(Read more...)

Even if the target, a disease on plants for example, is visible and easy to spot if one is up close and knows what to look for, there can still be a challenge in examining a large volume of targets if there are rows upon rows of plants either inside a greenhouse or outside in a field. Access can also be a problem, where a drone carrying a camera might work better.