| As the visual measurement technology is characterized with high precision,high efficiency,large amount of information and non-contact advantages,it has been widely used in the aerospace,automotive,electronics and other fields.With the improvement of the industrial manufacturing technology,the requirement of visual measurement accuracy is improved.Besides the image resolution directly affects the accuracy of the key feature information extraction and the final measurement.Therefore,the demand for high resolution images is becoming more and more urgent.However,due to the constraints of the optical sensor manufacturing process and cost,it is difficult to further improve the image acquisition hardware equipment,so we need to increase the image resolution by software.Image super-resolution reconstruction has the advantages of low cost and high quality and it can improve the image resolution on the basis of existing hardware devices,which shows a good potential application.Super-resolution reconstruction is mainly used to improve the visual effect of the image and obtain a clearer image on the subjective judgment currently but there is no application in the image feature extraction and visual measurement.Therefore,for the actual measurement requirements,this paper constructs the aided feature in the image firstly and reconstructs the feature information area by using the super-resolution reconstruction,so as to improve the precision of the image feature extraction and the visual measurement.The main research contents are as follows:(1)According to the requirement of aided feature in visual measurement and super-resolution reconstruction,this paper uses multi line laser to construct the feature.We analyze the projection and its adjustment of aided targets and at the same time take the compactness,practicability and reliability into consideration.A set of the laser array adjustment device is designed and processed,and the corresponding control system is developed by using Lab VIEW.The experiment results show that the device and system has the following features: stable working,high efficiency and good robustness,which can effectively meet the aided structure needs in the measurement and reconstruction.(2)Aimed at complex shape object or much key feature information and other situations that need for large-scale reconstruction of the image,focusing on the efficiency,flexibility and adaptability,a super-resolution reconstruction method based on area coverage principle is proposed.The registration of image achieved with aided targets and the image sequence is obtained by image inverse transformation.Finally,the initial high-resolution image estimation is corrected according to the size of the displacement to realize the super-resolution reconstruction.The simulation experiment shows that the average extraction error is reduced from 0.704 pixel to 0.33 pixel.The calibration plate angular point extraction results show that the mean square error is reduced by 46.6% and 47.5% in the horizontal and vertical directions respectively.The mean square error of the straight line fit is reduced by 31.7% after reconstruction in laser intersection extraction experiment.(3)Aimed at the measured object with simple construction and improving the reconstruction accuracy of the local key information,an adaptive dynamic threshold convex projection method is proposed.The correspondence between image pixel and the actual physics spatial parameters are solved with aided targets.The sequence images with sub-pixel displacement deviation are collected by mobile camera which is driven by micro-displacement platform.The super-resolution reconstruction is achieved by correcting the pixel of initial estimation according to the constraints of data consistency and amplitude.The resolution chart reconstruction experiment shows that the image detail has obvious improvement.The standard aluminum plate measurement results show that the measurement error is reduced from 0.52% to 0.18% after reconstruction. |
PDF Full Text Request