Ferrographic image multi-information fusing method based on D-S evidence theory

Abstract

The invention discloses a ferrographic image multi-information fusing method based on a D-S evidence theory. In order to solve the technical problem that heterogeneous information in ferrographic image wear particle recognizing is low in comprehensive utilization rate, binary image filtering and colored image filtering are achieved after wear particle image binaryzation treatment; the form features and color features of a wear particle image are synchronously extracted, hard output of a support vector machine is converted into probability output through a sigmoid function, and a probability distribution function is constructed for the fusing of the heterogeneous information of the form features and the color features of a ferrographic image. Multi-information fusing is achieved through theD-S evidence theory, the advantages of the form features sensitive to cutting and oxide wear particles and the color features sensitive to slide wear particles are combined, and the effective distinguishing of three types of failure wear particles is achieved. Compared with independently used color features and form features, the recognizing accuracy of the method is improved by 16.6% or above, the comprehensive utilization rate of wear particle information is effectively improved, and a new concept is provided for device wear monitoring.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis of mechanical equipment, and in particular relates to a multi-information fusion method of ferrographic images based on D-S evidence theory. Background technique [0002] Ferrography analysis technology is a kind of wear particle analysis technology that appeared in the field of international tribology in the 1970s. Digital image processing technology originated in the 1950s and has developed rapidly in decades. The beneficial combination of the two has further promoted the development of ferrography analysis technology. By analyzing the ferrographic wear particle image, the wear information of the equipment can be obtained, and the wear type and severity of the equipment can be predicted. Therefore, ferrographic wear particle analysis based on image processing technology has received extensive attention and research. [0003] At present, in the field of ferrographic wear particle image...

AI Technical Summary

Problems solved by technology

Although this ferrographic image processing method is simple and easy to implement, it lacks a finer filtering process after binarization, resulting in more impurity information mixed into the segmented binary image, which is not conducive to the subsequent wear particle feature extraction and analysis. Wear Particle Identification

Images