Spectronon™: Free Hyperspectral Imaging Software

Hyperspectral imaging creates digital images with much higher spectral, or color, resolution than traditional color imaging. If you are new to the concept of hyperspectral imaging, you might want to read our Introduction to Hyperspectral Imaging. A hyperspectral datacube (the 3-dimensional image file that is created by a hyperspectral camera) provides a detailed spectral curve for every pixel in the image. The high spectral resolution data enables differentiation between materials more accurately than is possible with color images, expanding possibilities in a wide range of applications (e.g., environmental monitoring, precision farming, sorting, quality control, diagnostics, etc.).

To work with the large quantity of hyperspectral data contained within a datacube, Resonon developed a proprietary software called Spectronon. Since Spectronon’s creation in 2005, we have been continually refining, improving, and enhancing its capabilities. Since that time, many thousands of people have downloaded Spectronon.

Spectronon’s visualization capabilities allow you to see hyperspectral images in many ways to facilitate interpretation of the spectral data. For example, Spectronon makes it easy to see the spectrum of a single pixel or a user-selected group of pixels.

Below, in Figure 1, the small gray boxes inside the green and blue circles indicate selections of pixels from two different types of candy that are nearly identical to the human eye. The mean value and standard deviation of the spectra from these pixels are plotted in green and blue in the upper right-hand corner of the Spectronon display, as indicated by the green and blue arrows.

That plot (enlarged in Figure 2) displays the wavelength vs. the signal of the pixels selected. One can easily see the spectral difference between the two types of red candies.

That plot (enlarged in Figure 2) displays the wavelength vs. the signal of the pixels selected. One can easily see the spectral difference between the two types of red candies.

Spectronon comes with many standard algorithms for classification and regression of hyperspectral data. Classification might be used to determine a good product from a bad product. Regression is best suited to monitor changes, such as how much an item has dried or how long it has been cooked.

Figure 3, below, shows an example classification map of hyperspectral imaging data using an algorithm known as the Spectral Angle Mapper (SAM). This algorithm compares every pixel in the image to the blue and green spectra shown in the plot. Pixels with spectra that are sufficiently similar to the blue spectrum are colored blue, and those that are sufficiently similar to the green spectrum are colored green. Pixels that are not similar to either spectrum are colored black. The classification map clearly shows that there are two types of red candy, with one type organized in the shape of an “I.” None of the yellow candy pixels have spectra similar to either type of red candy, so pixels of the yellow candy are colored black. The thresholds for similarity in the SAM algorithm are easily adjusted with the sliders circled in red in Figure 3. Results from other built-in algorithms (Logistic Regression, Random Forest, Euclidean distance, to name a few) can be quickly compared to determine which algorithm provides the best results for your particular application.

Spectronon is designed for ease of use. There is a quick start guide to get the user up and running. There is also a comprehensive user manual that includes clear descriptions of everything from basic to advanced capabilities and the details behind all the algorithms. Introductory tutorial videos to get you up and running are available on our Library page.

Spectronon has the ability to perform batch processing and data pre-processing steps. The user can also write their own plugins (in Python), essential for those looking to combine their own analysis algorithms with the power of Spectronon. Spectral plots can be saved as text files for easy import into Excel, Python, or similar software for further analysis. Images can be easily exported into many different file formats.

Download Spectronon and sample hyperspectral imaging datacubes to visualize, analyze, and gain a better understanding of the power of hyperspectral imaging.

For more details about Spectronon, see the User Manual for Spectronon and the Tutorial Videos on our Library page.

If you want to collect your own hyperspectral data, Resonon offers a variety of different hyperspectral imagers and hyperspectral systems to fit your particular application.