Figure 4 - uploaded by Karsten Lambers
Content may be subject to copyright.
Left: A given set W = {S1, S2, .., S6} of linear segments around a candidate point p0. Right: A graph G w for the set of linear segments. We assume an angle tolerance α such that all angle constraints are satisfied. Several node pairs of the graph are not connected by an edge due to the convexity constraint, which is not satisfied for an assumed convexity tolerance t. The red nodes of the graph are the nodes of the optimal maximal clique G c opt. The corresponding valid configuration Copt is marked in red on the left figure.
Source publication
We develop an approach for detection of ruins of livestock enclosures in alpine areas captured by high-resolution remotely sensed images. These structures are usually of approximately rectangular shape and appear in images as faint fragmented contours in complex background. We address this problem by introducing a new rectangularity feature that qu...
Contexts in source publication
Context 1
... smaller than the set of graph cliques K(G w ), the number of times the rectangularity measure ρ needs to be evaluated in Eq. (10) is considerably reduced in compari- son to Eq. (8). Since, in addition, there are efficient algo- rithms for the search of maximal cliques [2], computing the rectangularity feature is not computationally demanding. Fig. 4 (left) shows an example of a given set W = {S 1 , S 2 , .., S 6 } of linear segments and the optimal config- uration C opt = {S 1 , S 2 , S 3 , S 5 } in red, while Fig. 4 (right) shows the corresponding graph G w and the optimal max- imal clique G c opt in red. There are two additional max- imal cliques G c 1 and G c 2 and corresponding valid ...
Context 2
... to Eq. (8). Since, in addition, there are efficient algo- rithms for the search of maximal cliques [2], computing the rectangularity feature is not computationally demanding. Fig. 4 (left) shows an example of a given set W = {S 1 , S 2 , .., S 6 } of linear segments and the optimal config- uration C opt = {S 1 , S 2 , S 3 , S 5 } in red, while Fig. 4 (right) shows the corresponding graph G w and the optimal max- imal clique G c opt in red. There are two additional max- imal cliques G c 1 and G c 2 and corresponding valid config- urations C 1 = {S 2 , S 3 , S 4 , S 6 }, C 2 = {S 1 , S 2 , S 3 , S 4 }. They, however, have lower rectangularity values ρ(G c 1 ...
Similar publications
Debris disks are usually detected through the infrared excess over the photospheric level of their host star. The most favorable stars for disk detection are those with spectral types between A and K, while the statistics for debris disks detected around low-mass M-Type stars is very low, either because they are rare or because they are more diffic...
Citations
... In Figure 1 it is shown how the main streets and pathways of the Italian city of Marcianise, in the province of Caserta, match the same organization of Roman Centuriation. The automatic recognition of these remains is a relevant and challenging activity in the field of smart archaeology [16]. In fact, nowadays the study of such archaeological remains is still made manually and it is not trivial because original divisions are Moreover, there exist variations to the original schema that use, for example, rectangular patterns whose edge measures half an actus (or its multiple). ...
... In this paper we focus on the next phase that processes the results of deep learning elaboration to estimate the degree of Centuriation trough the identification of land-use patterns and/or regular shapes. In [16] authors developed an approach for the detection of ruins of livestock enclosures in alpine areas estimating a rectangularity feature that quantifies the degree of alignment of an optimal subset of extracted linear segments with a contour of rectangular shape. In [1] an image analysis procedure utilizing Radon transforms has been used to disclose landscape divisions relating to the Roman settlement and land use obliterated by agricultural intensification. ...
Methodologies and technologies for uncovering archaeological ruins and objects have been advanced by the utilization of Artificial Intelligence (AI). AI supports archaeologists to discover remains that are difficult to be identified manually as they must be sought on a vast territory or because they are hidden from direct observation. Here we present a methodology that integrates deep learning, computer vision and genetic algorithm techniques to identify, from large aerial pictures, remains of the Centuriation, which is an ancient Roman system for the division of lands.
... In Figure 1 it is shown how main streets and pathways of the Italian city of Marcianise, in province of Caserta, match the same organization of Roman Centuriation. The automatic recognition of these remains is a relevant and challenging activity in the field of smart archaeology [6]. In fact, nowadays the study of such archaeological remains is still made manually and it is not trivial because original division are partially lost. ...
... While no incontrovertible evidence of centuriation was located, satellite data proved to be of use in surveying medium-scale rural patterns. In [6] authors developed an approach for detection of ruins of livestock enclosures in alpine areas captured by high-resolution remotely sensed images. These are structures usually of approximately rectangular shape and appear in images as faint fragmented contours in complex background. ...
Archaeological research that uncovers ancient artifacts is of critical importance, since ruins and objects are the only way to deal with the history of a city. Artificial intelligence techniques can support the archaeologists to discover remains which are difficult to be identified manually as they must be sought on a vast territory or because they are hidden from direct observation. Here we present an original methodology that integrates deep learning and computer vision techniques to identify, from aerial pictures, remains of the Centuriation, that is an ancient Roman system for the division of the territory.
The applications of automated object detection in remote sensing archaeology have grown considerably in the last few years. This reading list has been compiled as a contribution to consolidating current perspectives at September 2016, and in support of the preceding paper on the broader issues of human and computer vision in archaeological prospection (Traviglia et al.).
(Full text available at https://hdl.handle.net/1887/3156937)
We develop an approach for the detection of ruins of livestock enclosures (LEs) in alpine areas captured by high-resolution remotely sensed images. These structures are usually of approximately rectangular shape and appear in images as faint fragmented contours in complex background. We address this problem by introducing a rectangularity feature that quantifies the degree of alignment of an optimal subset of extracted linear segments with a contour of rectangular shape. The rectangularity feature has high values not only for perfectly regular enclosures but also for ruined ones with distorted angles, fragmented walls, or even a completely missing wall. Furthermore, it has a zero value for spurious structures with less than three sides of a perceivable rectangle. We show how the detection performance can be improved by learning a linear combination of the rectangularity and size features from just a few available representative examples and a large number of negatives. Our approach allowed detection of enclosures in the Silvretta Alps that were previously unknown. A comparative performance analysis is provided. Among other features, our comparison includes the state-of-the-art features that were generated by pretrained deep convolutional neural networks (CNNs). The deep CNN features, although learned from a very different type of images, provided the basic ability to capture the visual concept of the LEs. However, our handcrafted rectangularity-size features showed considerably higher performance.
