SPKTRA ApS – Vision Control

SPKTRA InSPEK - Neural Network software

InSpek Software Solution

InSpek consists of three main components: 

Analysis

The analysis tab is for inspecting and understanding acquired hypercubes. This tab is intended for professionals wanting to investigate reflectance across wavelengths and compare between different hypercubes. 

Build

The build tab is for anyone wanting to create a vision solution. 

Simply mark which areas are to be separated and INSpek’s machine learning will segment images according to the marked categories.
INSpek’s machine learning algorithm works by comparing all pixels in a new hypercube to pixels in the marked areas. INSpek’s algorithm is capable of using the full spectrum of all pixels to determine which category any pixel is most likely to belong to.  

Deploy

The deploy tab is intended to deploy models created in the “build” tab. In this tab models can be broadcasted to predefined configurations of camera and light modules. Thus, models can be deployed to an arbitrary amount of cameras and light modules in different configurations. 

Running deployments can also be monitored and changed if new requirements or models become available. 

InSpek is intended as an all-in-one software package that handles everything related to our hyperspectral cameras while allowing the operator to focus solely on marking what is desired to separate and monitor the health of running deployments.