Research On The Denoising And The Restoration Of Frame Interpolation Of X-band Nautical Radar Sea-surface Image

The real-time ocean surface parameters can be obtained by the sea image data which is got through X-band nautical radar remote sensing survey, and the sea image data also provides accurate data basis for the later inversion. But there are many problems in the image data of sea surface which is collected by the X-band nautical radar previously, such as the noise and the data lost of a frame, and these problems directly affect the inversion of many ocean surface parameters. To overcome those problems, this paper has put forward research on the pretreatment of original image data, which prepares for the later inversion of many ocean surface parameters.Firstly, the traditional denoising methods such as mean filtering method and median filtering method are not suitable for radar image denosing for the difference in space distribution structure between radar image and traditional image. The Inverse Distance Weighted Algorithm (IDWA) is based on the space-time characteristics of the X-band radar sea surface image. First the noise in radial pattern can be separated by line detection through this algorithm. Considering the special pattern of the noise, a method of threshold segmentation has been put forward. Furthermore, interpolation in polar coordinates has been adopted on these noise points. The IDWA improves on both subjective vision and objective quality compared with traditional denoising methods. But there is shortcoming of this method, which is that the denominator may be zero during the process of interpolation, resulting in the black spots appearing in the image, which will influence the inversion results. This paper put forward an improved algorithm to overcome this problem and achieved good objective and subjective results.Secondly, this paper provides a method of image restoration based on motion estimation and motion compensation of video sequence restoration technology when a frame of successive image sequence lost. Search algorithm is the improved three-step motion search algorithm combined with bi-direction motion estimation according to radar image pixel motion characteristics to get the vector field, and the vector field is processed later. At last the missing frame can be reconstructed. In order to improve the accuracy of image restoration, this paper also quotes the sub-pixel motion estimation to improve the accuracy of the vector field and experiment results show that the algorithm can obtain good recovery quality.