Dual View X-ray Inspection System for Foreign Objects Detection in Canned Food

摘要

X-ray inspection technique for foreign objects in food products can determine and mark the presence of contaminants within the product by using image processing and pattern recognition technique on the X-ray transmission images. This paper presents the dual view X-ray inspection technique for foreign objects in food jar via analyzing the weak points of the traditional single view X-ray inspection technique. In addition, a prototype with the new technique is developed in accordance with glass splinters detection within the food jar (glass jar especially) which is a typical tickler. Some algorithms such as: adaptive image segmentation based on contour tracking, nonlinear arctan function transform and etc., are applied to improve image quality and achieve effective inspection results. The false recognition rate is effectively reduced and the detection sensitivity is highly enhanced. Finally the actual test results of this prototype are given.