Lets take a look at how Stream uses multi-spectral cameras, spectroscopy and machine learning to change the world of detection.
To understand how an invisible element can be detected using a spectral camera, it’s helpful to know that photons bounce off every plant, every nutrient, every disease in a unique way which identifies the specific particle.
Sometimes the wavelength is in the visible spectral range and is visible to the human eye. Light in the infrared range (known as the IR) is not visible, but it can be detected with a spectral camera.
Today we use cameras to capture these different wave lengths and send the data to our artificial intelligence for analysis.
The first question you have to ask yourself, is what item or object do you need to detect?
Example of Detecting Organic vs Polyethylene Leaves
In this example of classification, we are looking to decipher the difference between Organic and Polyethylene leaves.
We begin by selecting the objects that we would like to identify. On the right, there is a picture of 3 different leaves with 2 of them being Polyethylene. We are looking to identify the Polyethylene leaves.
Proceed to the Stream Analytics Site to log in. Here you can create a free account where you can have your results analyzed.
With a spectral camera, we are able to take a picture, or we can use a spectrometer to take a probe of a sample simply by touching it.
The image or data is then sent to the Stream Analytics Engine, where you should be signed in to your free account. The Stream Analytic Engine will use an Analytics App (Click here for more info) to identify the elements in your image or data.
In this case we will be using the Real and Fake Leaves Analytics App.
If a multi-spectral camera is used to capture the object, the detected pixels are false colored in the photograph. With spectrometer data the test results are represented with a percentage or graph.
In this example the Polyethylene leaves have been colored red, while the Organic leaf has been colored blue. The background has been colored green.