Spectral band identification tool for hyperspectral imaging-based diagnostics.

Stream Technologies是一家开发高光谱成像设备的加拿大公司。当一个 用传统的照相机拍摄对象的图像，图像中每个像素的强度是 表示图像中给定点处的所有波长（颜色）的光的强度。与 在高光谱成像中，该图像被分割成一组空间上相同的图像;然而，每个图像处的信号 每个图像的像素基于特定波长范围内的光强度（颜色）。以来 不同的材料吸收和反射/发射不同波长的光，这取决于它们的化学性质， 不同的图像可以提供关于图像的不同区域的空间分辨的化学信息。 挑战在于确定如何调整系统，以便收集适当的光谱带 使得可以从数据集中收集期望的信息。该研究项目将开发一个 实验协议，以收集调整成像系统所需的数据并对所得数据进行建模 以便将来可以获得期望的信息。最初的工作将与模型系统（例如， 混合在咖啡豆中的鹅卵石，识别不同的糖果等）。一旦该协议建立，更多 将探索具有挑战性的系统，包括区分芥末和油菜籽，测量肉类， 水分/蛋白质/脂肪等。

Stream Technologies is a Canadian company that has developed hyperspectral imaging equipment. When one takes an image of a subject with a conventional camera, the intensity of each pixel in the image is representative of the intensity of all wavelengths (colours) of light at a given point in the image. With hyperspectral imaging, this image is split into a suite of spatially identical images; however the signal at each pixel of each image is based on the intensity of light across a specific range of wavelengths (colours). Since different materials absorb and reflect/emit light at different wavelengths, depending on their chemistry, the different images can provide spatially-resolved chemical information about different regions of the image. The challenge lies in determining how to tune the system so that appropriate spectral bands will be collected such that the desired information can be gleaned from the data set. This research project will develop an experimental protocol to collect the data necessary to tune the imaging system and model the resulting data such that the desired information can be obtained in the future. Initial work will be with model systems (e.g. pebbles mixed in coffee beans, identifying different candies, etc). Once this protocol is established, more challenging systems will be explored including distinguishing mustard and canola seeds, measuring meat moisture/protein/fat, etc.