Search

This paper proposes a line-matching algorithm based on feature grouping and a line band descriptor (LBD) to address the insufficient reliability of individual line descriptors for line matching. First, the algorithm generates line-pairs according to geometrical relationships such as the distances and angles between line segments extracted from a single image. Subsequently, the algorithm employs the epipolar line of intersection between two lines in a reference line-pair to constrain candidate pairs corresponding to the reference line-pair. Thereafter, each line in the reference line-pair is considered individually, and its support region and the corresponding support region of each candidate line in the candidate pairs are established, following which an affine transformation is used for unifying the sizes of the reference support region and the candidate support region. Moreover, the LBD descriptor is then used for describing the reference and candidate lines. The Euclidean distances between the reference line and each candidate line descriptors are calculated, and the nearest neighbor distance ratio (NNDR) is used as a criterion for determining the final matching. Finally, the one-to-many and many-to-one line correspondences in matching results are transformed into one-to-one line correspondences by fitting multiple lines to the new line; simultaneously, incorrect matches are eliminated. The experimental results show that the proposed algorithm yields reliable line-matching performance for close-range images.

A new straight line matching method for aerial images is proposed in this paper. Compared to previous works, similarity constraints combining radiometric information in image and geometry attributes in object plane are employed in these methods. Firstly, initial candidate lines and the elevation values of lines projection plane are determined by corresponding points in neighborhoods of reference lines. Secondly, project reference line and candidate lines back forward onto the plane, and then similarity measure constraints are enforced to reduce the number of candidates and to determine the finial corresponding lines in a hierarchical way. Thirdly, "one-to-many" and "many-to-one" matching results are transformed into "one-to-one" by merging many lines into the new one, and the errors are eliminated simultaneously. Finally, endpoints of corresponding lines are detected by line expansion process combing with "image-object-image" mapping mode. Experimental results show that the proposed algorithm can be able to obtain reliable line matching results for aerial images.

Geographic information sciences (GIS) have been entering an era of information explosion. The data-related geographic can be divided into many classes, according to their sources and format, such as raster dataset, shape file, textual information, and voice. Locality description, which is a common form of voice, conveys considerable spatial information and can be derived from our daily communication. The issue of dealing with the locality description information is a research hot spot of next-generation GIS for many scholars.Locality description reflects direct or indirect human interaction with environment directly. As an external expression of cognition, the uncertainty that is associated with locality description is inevitable. Locality description generally contains spatial relationships (i.e., topological, distance, and direction relations) and reference objects (ROs). Any feature with a name can be regarded as an RO. Topological relations, which convey rough information-related locality and can be refined by distance or direction relations, are seldom used directly in locality description positioning. The distance and direction relations are usually combined to describe locality, which conveys many clues to position locality.Humans have a weak sense of direction indoors, and relative directions are used frequently in locality description. For example, locality description indoors can be given as follows: “Object A is in front of me, and object B is on my left. Context is an unavoidable topic of Locality description. The locality description is complex, either explicitly or implicitly, especially in a landmark reference system (i.e., a reference system where people can describe his locality with one or several landmarks), in which the nearest landmark can be selected easily to describe locality. On the basis of this context, the locality description (“Object A is in front of me, and object B is on my left”) stated above in an indoor landmark reference system (ILRS) implies that objects A and B are near the individual. Hence, the meaning of “Object A is in front of me, and object B is on my left” in ILRS is the same as that of “Object A is in front of me, object B is on my left, and they are all near to me”.This paper introduces a novel method of positioning localities indoors by using locality description in ILRS. To achieve positioning of localities with directions description and context in ILRS, we propose a joint probability function that consists of qualitative distance (i.e., near relation) and relative direction membership function. The qualitative distance membership function that considers both minimum Euclidean distance and the stolen area is based on fuzzy set. For consistency with cognition, some definitions are provided during the calculation of relative direction, which can also reduce the number of points to be explored from an algorithmic point of view.

In order to achieve more robust pose tracking and mapping of visual SLAM, the robotics researcher has recently shown a growing interest in utilising multiple camera, which is able to provide more sufficient observations to fulfil the frame registration and map updating tasks. This implies that better pose tracking robustness can be achieved by extending monocular visual SLAM to utilise measurements from multiple cameras.[1] proposed a visual SLAM method using multiple RGB-D cameras, which integrate the observations from multi-camera for camera tracking. However, they ignored the time-drift between the frames obtained by different cameras, which may result at inaccurate positions of observation used for map updating. Besides, loop closure detection was not been implemented. [2] constructed a multiple RGB-D system with three Kinects V2 camera. This work mainly concentrated on the intrinsic and extrinsic calibration and verify the effectiveness of mapping using multiply RGB-D cameras.