No abstract

Precision size measurements are widely used in industrial production such as candies and cookies in the food industry. However, it is a challenge to measure irregular objects in a universal device because of poor adaptability and low precision. In this paper, a high-precision measurement device for irregular objects based on machine vision is designed. It consists of a rotational adjustment platform, two imaging units, a weighing device and a host computer. Armed with this device, a novel algorithm with good robustness is proposed to rapidly measure irregular objects. The sub-pixel location algorithm based on spatial moment is used for the accuracy size measurement. Experiments show that the measurement accuracy is less than 25 m.