HAIRIS: A Method for Automatic Image Registration Through Histogram-Based Image Segmentation:

HAIRIS: A Method for Automatic Image Registration Through Histogram-Based Image Segmentation:

Abstract:

           

                Automatic image registration is still an actual challenge in several fields. Although several methods for automatic image registration have been proposed in the last few years, it is still far from a broad use in several applications, such as in remote sensing. In this paper, a method for automatic image registration through histogram-based image segmentation (HAIRIS) is proposed. This new approach mainly consists in combining several segmentations of the pair of images to be registered, according to a relaxation parameter on the histogram modes delineation (which itself is a new approach), followed by a consistent characterization of the extracted objects—through the objects area, ratio between the axis of the adjust ellipse, perimeter and fractal dimension—and a robust statistical based procedure for objects matching. The application of the proposed methodology is

illustrated to simulated rotation and translation. The first dataset consists in a photograph and a rotated and shifted version of the same photograph, with different levels of added noise. It was also applied to a pair of satellite images with different spectral content and simulated translation, and to real remote sensing examples

comprising different viewing angles, different acquisition dates and different sensors. An accuracy below 1° for rotation and at the subpixel level for translation were obtained, for the most part of the considered situations. HAIRIS allows for the registration of pairs of images (multitemporal and multisensor) with differences

in rotation and translation, with small differences in the spectral content, leading to a subpixel accuracy.

Introduction:

               Automatic image registration (AIR) is still a present challenge regarding image processing related applications. Remote sensing applications is one of the fields where further research on AIR methods is required.

               Image segmentation (IS) is still an actual field of research, regarding automatic methods of image processing. IS is generally defined as the process that partitions an image into regions, each of them fulfilling a given criteria, which can be from the image domain and/or feature space. From image segmentation methods, we expect the extraction of a set of objects present on an image, as we visually detect them.

               In other words, it is expected that a segmentation method acts as artifical intelligence on the identification of objects on a scene. However, the objective of the segmentation may be quite subjective, depending upon the detail and features we are expecting. For instance, on the segmentation of an image of a human body, one may be interested in delineating the whole body as a single object, or its constituent parts, which may become itself quite subjective.

             

Existing System:

          

             The rigid-body model under the scope of automatic image registration methods is still a present subject of research in particular under the scope of remote sensing applications  The problem of registering remote sensing images can roughly be locally seen as the determination of translations and a small rotation. Under the scope of computer vision applications, the rigid-body transformation may seem a simple problem to solve with many existing methods. However, under the scope of remote sensing applications, one of the major problems is

related to the radiometric content (due to multisensor or multispectral pairs of images).

           

Proposed System:

          

              In this paper, a method for automatic image registration through histogram-based image segmentation (HAIRIS) is proposed, which allows for a more detailed histogram-based segmentation, rather than the traditional methods, and consequently to an accurate image registration. HAIRIS is able to estimate the rotation and/or translation between two images— which may be multitemporal or multisensor—with small differences in the spectral content.This methodology begins with a preprocessing stage in order to reduce unnecessary detail on the images content, important for the subsequent histogram-based image segmentation phase (which includes a

relaxation parameter ). The objects extracted from the segmentation stage are characterized and matched according to some related properties, which finally allows for the statistically-based rotation and translation parameters estimation. The proposed method is based upon detecting closed similar regions in both images. Taking into account that the pair of images to be registered presented limited differences regarding their spectral content, it will usually be possible to detect similar regions in both images, even for regions with low contrast. Furthermore, one important characteristic of HAIRIS is the segmentation which is produced at different

levels, by considering a range of values for the relaxation parameter . This allows for the obtention of “several segmentations” and consequently to a more robust subsequent stage of initial matching. Moreover, it should be noticed that a closed region is a subjective concept. For instance, the segmentation of a

river may be seen as a line, but may also be considered a closed region depending upon its width.

Software Specification:

                  

                

COMPONENT	REQUIREMENTS
Front End	C#.NET
Database	MS-SQL Server 2000
Application Server	IIS
Operating System	Windows XP
Browser	IE

              

Hardware Specification:

 

HARDWARE	CONFIGURATIONS
No of system	Minimum 3
Processor	Intel CORE2Duo
Clock Speed	1 GHz or above
Cache Memory	2 GB
Base Memory	1 GB
RAM	1 GB
Hard Disk	80 GB
Floppy Disk Drive	1.44 MB
CD-ROM	䦋㌌㏒㧀좈໱琰茞ᓀ㵂Ü
Key Board	104 Keys
Display Device	15” Color Monitor

Conclusion:

          

               In several applications, the registration model only assumes rotation and translation. In this paper,

a new approach for automatic image registration through histogram- based image segmentation (HAIRIS) is proposed, with clear advantages by joining these two main areas of image processing. With the filtering step—an important preprocessing stage of the proposed methodology—the objective is to transform the original image in order to take advantage of the psychophysical aspects of the human visual system. The Wiener filter is

one of the solutions among several other possible alternatives. As a consequence, beyond the intended detail reduction, the distinction between objects on the image is sometimes lost, since it induces a significant smoothing on the entire image, including the objects limits. The main drawback of HAIRIS is the computational time, mainly associated to the segmentation stage. However, it is expected to optimize the implementation code in the future in order to provide a faster performance. In this work, HAIRIS was applied to single-band images at a time. However, in the future, adequate transformations (such as principal component analysis, independent component analysis, among others) of multi- (or hyper-) spectral images to singleband images will certainly lead to even better results, rather than using the information of a single spectral band. Furthermore, under the scope of applications with images having less evident objects, as is the case of remote sensing images, HAIRIS has shown to correctly register a pair of images at the subpixel level covering a wide range of situations (including multitemporal and multisensor).