Research On The Detection Method Of Strip Open Window Pipe Based On Thermal Imaging

Oil screen has very important role in the process of oil gathering,its quality directly affect the purity of collected oil,oil screen wear detection precision and the improvement of automation level has always been a hot research direction.At present,the common detection methods are mainly manual detection and magnetic flux leakage detection,but there are problems such as insufficient accuracy and poor real-time performance.This thesis is based on the limitations of the current oil slit screen detection as a starting point,designed a set of experimental detection system,to achieve the high precision detection of the oil screen.The main research contents of this thesis include:In terms of infrared image acquisition,because passive imaging detection in non-heating system has low imaging quality,active imaging detection is adopted which can better reflect the structural characteristics of the measured material.ECPT system with shorter heating time,better process controllability and higher imaging quality is adopted for heating.In the aspect of image noise reduction,considering the traditional single template spatial neighborhood smoothing filtering algorithm in the process of image to noise may be the image edge fuzzy problem,on the improvement of improved algorithm can suppress the noise,good neighborhood smoothing filter algorithm and optimize the space damage to the image edge details,good to protect the image detail.In the aspect of image enhancement,in view of the existing single enhancement algorithm is difficult to both enhance the image contrast and highlight the problem of image edge,in this thesis,based on the guided filter,this thesis proposes a layered image enhancement improvement method,using the BBHE algorithm instead of HE and PE algorithm,based on hierarchical enhancing to improve image contrast of the image,Improve the overall average brightness of the image,and use Gamma transform instead of linear transform to highlight the details.This algorithm can improve the objective evaluation indexes such as information entropy,average gradient and standard difference,and can be applied to the thermal image detection of the strip window of oil screen.In the aspect of edge extraction and fitting,in order to solve the problem of incomplete and inaccurate linear fitting information of Hough transform and Radon transform,Canny operator combined with LSD algorithm is used to get more accurate edge skeleton information,and finally achieve the high-precision detection of oil slit screen.The simulation results show that the proposed method is superior to the traditional method in detection accuracy.