Identification of Different Food Grains by Extracting Colour and Structural Features using Image Segmentation Techniques

Objectives: This paper proposes a method of identifying the different food crops which is automated. Identification of food crops can be automated using image segmentation techniques which reduces manual work. Method: Images of various food grains like wheat, paddy, cereals, barley and many more are captured using high resolution camera. Grab Cut algorithm helps in extraction only the part of image that the user is interested in. The colour features and outer boundaries of the food grains are extracted using Watershed algorithm and Canny detect detection. Findings: Extracting only the required part of an image i.e. the food grains, have made the processing easier as less features will be extracted. This reduction becomes very important when large data-set is considered. Many of the techniques used for segmenting grains consider only one feature, either colour or structural feature. The proposed system uses both colour and structural features which makes segmenting more accurate. Canny edge detection and Watershed algorithm takes less time for computation when compared with other techniques. Applications: An electronic hardware component can be built and the software part which identifies the grains can be embedded with it and used in the agricultural fields for separating food grains.