Spectroscopy and Machine Vision Solutions
Artificial vision based on deep learning (neural networks) mimics, to a certain extent, the way in which the human brain processes visual information, so that, as with the brain, it is essential that there is first a learning process. To do this, a sufficiently representative set of images of the cases to be detected must be available, which serves to “train” the neural network, i.e. the network must be able to “interpret” or “recognise” the characteristics of interest. Once the constructed network has been validated, it is used to inspect each new image, the output of the network being a label that identifies that image as belonging to one of the pre-established classes. In practice, a neural network is a set of numerical values, collected in a digital file, specific to a particular task, so that dedicated software “applies” the network to each image in order to obtain the label corresponding to the class to which the image appears to belong.
Traditional machine vision algorithms are deterministic. This means that, by means of mathematical functions, whether generic or specific, each image is analyzed in search of the features of interest. This approach limits their application to very specific cases, where the features are very marked or evident with respect to the rest of the image elements, and always requires the intervention of human experts. In the real world, circumstances favorable to the use of deterministic algorithms are not at all frequent. On the contrary, there are many interfering elements that greatly affect the accuracy of the result. Deep learning, on the other hand, does not start from more or less realistic assumptions but, in the learning phase, automatically adapts to the circumstances, i.e. it is less sensitive to unforeseen variations, and can achieve this without the need for theoretical assumptions or the work of an expert.
The most frequent application is the detection of morphological defects in order to activate a rejection system that discards those products labelled by the neural network as defective. For example, detection of foreign bodies or defective units in foodstuffs, intolerable colour differences in parts, textiles or food products or the objective quantification of faults or defects in order to calculate a fair price for the various quality grades of a product, for example defects in grains for sale and marketing.
