École Polytechnique Fédérale de Lausanne (EPFL)
École Polytechnique Fédérale de Lausanne (EPFL)
École Polytechnique Fédérale de Lausanne (EPFL)
Cadaster plans are cornerstones for reconstructing dense representations of the history of the city (di Lenardo and Kaplan, 2015). They provide information about the city urban shape, enabling to reconstruct footprints of most important urban components (buildings, streets, canals, bridges) as well as information about the urban population and city functions (census information, property, rent prices, etc.) (Noizet et al, 2013). Cadaster plans are usually the results of coordinated campaigns with standardized methods of measurement and representation. This means that large sets of documents follow the same representation conventions. This regularity opens the possibly of efficient automated process for analyzing them and possibly transforming the information they contain in geo-referenced databases that can be used as part of historical geographical information systems (Gregory et al, 2001).
However, as some of these handwritten documents are more than 200 years old, the establishment of a processing pipeline for interpreting them remains extremely challenging. This may explain why, to our knowledge, no such system exists in the literature. This article reports our effort in this domain, presenting the first implementation of a fully automated process capable of segmenting and interpreting Napoleonic Cadaster Maps of the Veneto Region dating from the beginning of the 19th century. Our system extracts the geometry of each of the drawn parcels, and classifies, reads and interprets the handwritten labels. We believe the general principle of technologies used in the process could be adapted to other cadastral funds, although this has not been tested in the present study.
Literature on map processing includes works on many different types of maps, from roads to topographic maps, including hydrographic and cadastral maps. Most studies focus on particular problems and features and thus develop techniques that are highly map specific (Chiang et al, 2014).
Our work addresses the particular case of the Napoleonic cadaster of Venice dated 1808, but aims at developing a method highly adaptable to other cadasters with little extra effort.
We propose a system that segments the cadastral map, identifies and extract segmented objects such as parcels and identifiers and recognizes the extracted hand-written digits. A demo code with examples of the results can be found at https://github.com/dhlab-epfl/cadasters.
The method is summarized in Fig. 1.
Figure 1: Overview of the system
Usually, the processed images are ancient documents that have been digitized. To deal with the natural ageing of paper and eventual spots on the map without losing details, we use a non-local means de-noising method (Buades et al, 2005) to smooth the image.
We address the task of extracting the desired information from the document as a segmentation problem, which is a recurrent problem in image processing. A graph-based segmentation approach is adopted, which models the image as a weighted undirected graph. This allows to process the pixels or regions in the spatial domain of the image but also to use higher level information such as connections, similarities and dependencies between the elements.
Because a group of pixels sharing some similarities are more perceptually meaningful than a simple pixel, we use SLIC method (Achanta et al, 2012) to create superpixels. Superpixels are clusters of pixels that share similarities and spatial proximity and have the advantage of reducing the complexity of image processing tasks.
A graph is a mathematical structure composed of vertices and edges, representing a system of connections or interrelations among a set of objects. It is widely used to model relations, to study information systems or to organize data. In our case, the graph representing the image is initialized with superpixels as vertices. Its edges connect neighboring vertices (superpixels) and each edge has a weight which is a measure of the dissimilarity between neighboring elements. The distance (or dissimilarity) metric is based on color and edge/ridge features.
The oversegmentation of the image resulting from superpixel generation is then reduced by grouping superpixels into homogeneous regions and merging the corresponding graph vertices. Our approach uses global homogeneity, meaning that the method minimizes intragroup dissimilarity and maximizes intergroup dissimilarity. The ‘dispersion' of edge weights (i.e standard deviation within a region) allows to spot high-weighted edges within a group and thus disconnect dissimilar vertices (i.e remove their edge) to end up with independent homogeneous regions.
The merged regions are classified into 3 classes: text, contour/delimitations and background (smooth textures such as parcels or streets) using a SVM classifier. The training data is composed of manually annotated samples of maps coming from the Napoleonic cadaster of Venice.
The classification results allow the determination possible parcels candidates. A flood fill algorithm is applied, using a ridge detector to indicate boundaries. The chosen ridge detector was originally developed as a vessel enhancement filter (Frangi et al, 1998) and looks for multiscale second order local structures of the image that can be considered tubular. The obtained measure indicates how similar the structure is to a tube, and so it is able to detect ridges. Starting from one point in the regions labeled as background (seed point), the flood fill algorithm floods each zone, i.e parcels, streets, etc. and stops at the boundaries (output of the ridge detector).
Each parcel of the image is extracted as a polygonal shape and the polygon's corner points are stored in GeoJSON format. If the image file is geo-referenced and contains geographical information (a GTIFF file for instance), polygons are exported according to the spatial reference system provided. This allows a fast and easy integration of the shapes into a geographic information system (GIS) and geographic information on the parcels can easily be collected.
The parcel identifier is usually contained within the parcel. This observation and the extracted polygons' information can be used to correct misclassified text regions and improve identifier extraction. Elements labeled as text regions are localized, delimited by bounding boxes and grouped so that neighboring characters are extracted together. Again, information from polygons is used to determine whether neighboring digits belong to the same identifier or not (i.e whether neighboring digits are located in the same parcel/polygon). Boxes that do not correspond to identifiers or digits are removed according to specific criteria. Finally, the boxes containing the parcels' identifiers are extracted.
Since the digit recognition step requires horizontally oriented digits to output accurate prediction, the identifiers' boxes are rotated. A principal analysis component is applied to the binary image of the extracted numbers to determine the angle of the rotation.
The horizontally oriented numbers are separated into digits that are processed individually. A good digit segmentation is primordial since connected or overlapping digits lead to incorrect recognition. A Convolutional Neural Network (CNN) with two convolutional layers, two fully connected layers and a final softmax layer for multiclass classification is used to predict the identifiers. The CNN is trained on a mixed dataset composed of MNIST dataset (LeCun et al, 1998) and digit samples from Sommarioni register and has a performance of 99.1%. When predicting the numbers, the network outputs the inferred number with a confidence level indicating the reliability of the result.
The proposed approach shows promising results in parcel extraction and identifier recognition. We performed the first ‘proof-of-concept' evaluations on manually labeled data taken from different cadaster samples. The total number of annotated objects are shown in Table 1.
Most parcels and identifiers were correctly extracted (Table 2 & 3), which assured us of the feasibility of their automatic extraction. The precision can still be increased for example by using feedback from digit recognition results, i.e, the prediction and its confidence level permit the discarding regions where no reliable identifier has been recognized.
Figure 2: Sample of results: (a) original image, (b) polygon approximation of parcels, (c) extracted parcels and (d) identifier localization
Parcels with labels
All parcels (with and without lab els)
Table 1: Count of ground-truth objects
Table 2: Results of parcel extraction with different Intersection over Union (IoU) thresholds
Concerning the digit recognition, only 10% of the identifiers had their digits correctly recognized. Since the models used have shown good performance on nicely detached digits, this is not the fault of the recognition algorithm itself but rather of the digit segmentation procedure. The current segmentation is the main hindrance to an efficient digit recognition, thus, further work should focus on a better number processing algorithm. Another alternative is to avoid the segmentation problem and use a recurrent neural network such as LTSM to process the number as a sequence.
Our work shows promising results for easing and accelerating cadaster processing, especially given our method's efficient parcel segmentation and digit identification. Moreover, the export of a parcel's geometry into GeoJSON format opens up further perspectives to efficiently geo-reference ancient maps. The system can be extended and integrated into a user interface to take better advantage from the results, for example by allowing the user to correct or add information about parcels and identifiers.
Table 3: Results of parcels' number localization with different Intersection (overlapping percentage) thresholds.
Table 4: Results of parcels' number recognition
The proposed method creates a bridge between previously seperate two data types: the raster object and the vector object. Currently, web-mapping tools consider vector objects as separate layers on the raster maps, and each object needs to be manually redesigned. The automatic vectorization process enables us to perform the visualization and annotation processes directly on the cartographic source without the prerequisite of complex skills. It should greatly facilitate large scale exploitation of such kinds of documents.
features,” in Document Analysis and Recognition, 1993., Proceedings of the Second Interna- tional Conference on, pp. 577-580, IEEE.
Chen, Y.-K. and Wang, J.-F. (2000). “Segmentation of
single-or multiple-touching handwritten numeral string using background and foreground analysis,”
Pattern Analysis and Machine Intelligence, IEEE
Transactions on, vol. 22, no. 11, pp. 1304-1317.
di Lenardo, I. and Kaplan, F. (2015) “Venice time machine: Recreating the density of the past,” in Digital Humanities 2015, no. EPFL-CONF-214895.
Noizet, H., Bove, B., and Costa, L. (2013) “Paris de parcelles en pixels.”
Gregory, I. N., Kemp, K. K., and Mostern, R. (2001). “Geographical information and historical research:
Current progress and future directions,” History and
Computing, vol. 13, no. 1, pp. 7-23.
Chiang, Y.-Y., Leyk, S., and Knoblock, C. A. (2014). “A survey of digital map processing techniques,” ACM
Computing Surveys (CSUR), vol. 47, no. 1, p. 1, 2014.
Buades, A., Coll, B., and Morel, J.-M. (2005). “A non-local
algorithm for image denoising,” in Computer Vision
and Pattern Recognition, 2005. CVPR 2005. IEEE
Computer Society Conference on, vol. 2, pp. 60-65,
Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., and Susstrunk, S. (2012). “Slic superpixels com- pared to state-of-the-art superpixel methods,” Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 34, no. 11, pp. 2274-2282.
Frangi, A. F., Niessen, W. J., Vincken, K. L., and Viergever, M. A. (1998). “Multiscale vessel
enhancement filtering,” in Medical Image Computing
and Computer-Assisted Interventa- tion—MICCAI'98,
pp. 130-137, Springer.
LeCun, Y., Cortes, C., and Burges, C. J. (1998). “The mnist database of handwritten digits”.
Strathy, N.W., C. Y. Suen, C.Y., and Krzyzak. A. (1993).
“Segmentation of handwritten digits using contour
If this content appears in violation of your intellectual property rights, or you see errors or omissions, please reach out to Scott B. Weingart to discuss removing or amending the materials.
Hosted at McGill University, Université de Montréal
Aug. 8, 2017 - Aug. 11, 2017
438 works by 962 authors indexed
Conference website: https://dh2017.adho.org/
Series: ADHO (12)