Locate Hazelnut Axis Based on Projection Gradient Statistics

The hazelnut should be cracked before it is sold. Presently, the mechanical crack is often adopted to process the hazelnut for food. In order to ensure the crack effect, it needs to manual located the axis of the hazelnuts so that the hazelnuts can be cracked along the axial direction. This paper proposes an innovative method for axial locating of hazelnut based on projection gradient statistics. Firstly the images of the hazelnut on the production line are obtained by industrial camera, and then the edge and contour of hazelnut are extracted according to the difference values of brightness between hazelnut image and the background area. Secondly, an adaptive circle is used to select the optimal projection area. The center of the adaptive circle is set on the center of gravity of the hazelnuts, and the adjustable diameter is analyzed also. The region selected by the adaptive circle in the hazelnut image is projected from different orientations. Finally, the sum of the gradients of each projection histogram is calculated and the axial orientation of the hazelnut is determined based on the statistical values. The experimental results show that the accuracy of the method for the axial recognition of the hazelnut is up to 94.1%, which satisfies the requirements for axial locating of hazelnut in automatic test system.