Ricard Prados

A Novel Blending Technique for Underwater Gigamosaicing

The fusion of several images of the same scene into a single and larger composite is known as photomosaicing. Unfortunately, the seams along image boundaries are often noticeable, due to photometrical and geometrical registration inaccuracies. Image blending is the merging step in which those artifacts are minimized. Processing bottlenecks and the lack of medium-specific processing tools have restricted underwater photomosaics to small areas despite the hundreds of thousands of square meters that modern surveys can cover. Large underwater photomosaics are increasingly in demand for the characterization of the seafloor for scientific purposes. Producing these mosaics is difficult due to the challenging nature of the underwater environment and the image acquisition conditions, including extreme depth, scattering and light attenuation phenomena, and difficulties in vehicle navigation and positioning. This paper proposes strategies and solutions to tackle the problems of very large underwater optical surveys (gigamosaics), presenting contributions in the image preprocessing, enhancing, and blending steps, resulting in an improved visual quality in the final photomosaic. A comprehensive review of the existing methods is also presented and discussed. Our approach is validated by a large optical survey of a deep-sea hydrothermal field, leading to a high-quality composite in excess of 5 Gpixel.

A motion compensated filtering approach to remove sunlight flicker in shallow water images

A common problem in video surveys in very shallow waters is the presence of strong light fluctuations, due to sun light refraction. Refracted sunlight casts fast moving patterns, which can significantly degrade the quality of the acquired data. Motivated by the growing need to improve the quality of shallow water imagery, we propose a method to remove sunlight patterns in video sequences. The method exploits the fact that video sequences allow several observations of the same area of the sea floor, over time. It is based on computing the image difference between a given reference frame and the temporal median of a registered set of neighboring images. A key observation is that this difference will have two components with separable spectral content. One is related to the illumination field (lower spatial frequencies) and the other to the registration error (higher frequencies). The illumination field, recovered by lowpass filtering, is used to correct the reference image. In addition to removing the sunflickering patterns, an important advantage of the approach is the ability to preserve the sharpness in corrected image, even in the presence of registration inaccuracies. The effectiveness of the method is illustrated in image sets acquired under strong camera motion containing non-rigid benthic structures. The results testify the good performance and generality of the approach.

Planar homography: accuracy analysis and applications

Projective homography sits at the heart of many problems in image registration. In addition to many methods for estimating the homography parameters (R.I. Hartley and A. Zisserman, 2000), analytical expressions to assess the accuracy of the transformation parameters have been proposed (A. Criminisi et al., 1999). We show that these expressions provide less accurate bounds than those based on the earlier results of Weng et al. (1989). The discrepancy becomes more critical in applications involving the integration of frame-to-frame homographies and their uncertainties, as in the reconstruction of terrain mosaics and the camera trajectory from flyover imagery. We demonstrate these issues through selected examples.

State of the Art in Image Blending Techniques

In this chapter the main state-of-the-art techniques are presented and described. There are three main groups of blending algorithms, each of them showing some benefits and drawbacks. On the one hand, transition smoothing methods minimize the visibility of the seams between two images fusing the image information of the common overlapping area. A drawback of this group of methods is that geometrical image misalignments and moving objects may cause the visualization of artifacts on the overlapping regions. On the other hand, optimal seam finding methods compute the optimal placement of the seam in order to minimize the photometric differences along the path. In the case of this group of methods, problems may appear when joining images acquired with changing illumination conditions or different time exposures. Finally, hybrid methods combine both strategies by fusing the image information around an optimally computed seam. This last group of methods allows avoiding the above mentioned problems. The chapter also proposes a classification of the methods of the literature based on their nature and capabilities. The aim of this classification is to discern the optimal strategy to blend large-scale high-resolution underwater photo-mosaics.

Underwater 2D Mosaicing

The current chapter describes the main steps involved in the photo-mosaic building process. These steps comprehend the geometrical registration and warping of the images into a single common reference frame, along with an estimation of the topology of the trajectory performed by the UV, and a global alignment of the recovered trajectory. A widely extended geometrical registration strategy consists of identifying common image features between the involved images, using different image feature detectors. These image features, once identified, become correspondences that are used to estimate the camera motion between consecutive images, as well as to perform a global alignment of the estimated trajectory. Global alignment of all the involved images allows providing geometrical consistence to the underwater map. At the end of the chapter the problems and issues of the photo-mosaicing process are pointed out, with the aim of demonstrating the relevance of image blending techniques as a final step of the photo-mosaicing process.