Research On Methods For Airport Runway Diseases Detection Based On Vision

In recent years,civil aviation technology has developed rapidly,and the security guarantee in the airport area has played a decisive role in the successful take-off and landing of aircraft.Airport runway disease brings huge safety hazards to aircraft flight safety.For the research on the visual detection method of airport runway disease,this paper mainly studies the image processing and the detection of disease detection during the collection of airport runway,which includes three aspects: research on design image acquisition system.and calibration methods of visual inspection system,research on image pre-processing before airport runway disease detection,airport runway disease detection research.(1)Research on the calibration method of linear CCD camera,To improve the calibration accuracy of the linear array CCD camera and determine the accurate camera mode.Here combining the imaging characteristics of the linear array CCD,the small hole imaging model and the vertical method are used to establish a non-parametric model of the linear array CCD camera.and the mapping relationship between spatial feature points and imaging points and the size of the distortion between the ideal imaging point and the actual imaging point are directly determined.As an evaluation criterion,finally analyzing its advantages.(2)Research on image pre-processing before airport runway disease detection.Firstly,error compensation is performed for the collection of airport runway images;Secondly,for the mixed noise of Gaussian noise and salt and pepper noise often appearing in the environment,based on the sensitivity of adaptive filtering algorithm to salt and pepper noise,the sensitivity of weighted gradient reciprocal filtering algorithm to Gaussian noise,and the advantages of bilateral filtering algorithm to preserve edge and reduce noise,A bilateral filtering algorithm based on adaptive gradient weighted reciprocal is presented.Then the above three algorithms to perform filtering test experiments on Gaussian noise,salt and pepper noise and mixed noise of different intensities are used to evaluate the filtering effect through the peak signal-to-noise ratio and average absolute error,and to prove the proposed image noise filtering algorithm feasibility and noise filtering advantages;Finally,in order to obtain the airport runway panorama,the image stitching algorithm is researched,and the effect of stitching test is analyzed.(3)Research on airport runway disease detection methods.Aiming at the detection of airport runway droppings,a detection algorithm for runway disease droppings based on wavelet transform and morphology is proposed;Aiming at the detection of cracks and glue marks on airport runways,a runway disease image segmentation using BP neural network combined with genetic algorithm is proposed;The experimental results prove the effectiveness of these two methods,and the effective detection of airport runway diseases has been achieved.