Foreign Object Detection of Sintering Transport Belt Based on CNN
CK. Xiao, Bei Sun, YL Wang, LD. Qiu
Abstract
Belt transportation is the most important transportation method in sintering production. Sintered products are often mixed with large gangue, ironware (grate bar, steel plate) leading to belt scratches, material plugging. In order to eliminate the influence of foreign objects on sintering production, this paper studies the detection of sinter foreign objects. The detection, classification and location of foreign objects are based on CNN, and non-maximum suppression is improved by considering the difference of image features between anchors. The detection model is proposed and the results indicate the effectiveness of the method.
Topics & Concepts
SinteringProduction (economics)GangueComputer scienceResearch ObjectBar (unit)Artificial intelligenceMetallurgyMaterials scienceGeologyBusinessOceanographyBusiness administrationEconomicsMacroeconomicsIndustrial Vision Systems and Defect DetectionImage and Object Detection TechniquesImage Processing and 3D Reconstruction