Portable Assistive system for Visually Impaired using Raspberry Pi

People with problems in vision find it difficult to recognize the provisional products in the supermarket. The products vary in shape, color, size and weight, which play an important role in recognition. The proposed system consists of a module which will work on image processing and a separate module which works on voice processing. An effective real time image processing technique has been used to extract the GLCM features from the captured image. Different algorithms have been trained and tested for classifying the input image. SVM has been selected which gave a classification accuracy of 89.6%. Proposed algorithm has been loaded into Raspberry Pi v3. Portability has been considered as the major objective of the work. A battery backup was provided which made the individual to carry the device in whatever place and can use at any time. A headset is provided to the user from which he/she can hear the audio output of the product name.