Fig 3 - available via license: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
Content may be subject to copyright.
Figure showing a set of 32 different symbols used for Dataset-P&ID. Symbol 1 to Symbol 25 are complex symbols as they are structurally very similar and are detected using a Complex Shape Extraction module. Remaining Symbol 26 to Symbol 32 are detected using the Basic Shape Extraction module.
Source publication
![Comparison of Digitize-P&ID with prior-art [14] on 12 real-world P&IDs](publication/354461183/figure/tbl2/AS:1065987419959299@1631162614850/Comparison-of-Digitize-P-ID-with-prior-art-14-on-12-real-world-P-IDs_Q320.jpg)
Digitization of scanned Piping and Instrumentation diagrams(P&ID), widely used in manufacturing or mechanical industries such as oil and gas over several decades, has become a critical bottleneck in dynamic inventory management and creation of smart P&IDs that are compatible with the latest CAD tools. Historically, P&ID sheets have been manually ge...
Contexts in source publication
Context 1
... only candidates for merging are the series of segments with opposite orientation and in close proximity with each other which are obtained using the DBSCAN [5] algorithm. -Basic Shape extraction module: Among various symbols used in P&ID sheets, certain symbols are composed of primitive shapes, such as rectangles and circles (Figure 3). Some of these symbols are differentiated via the texts written inside them. ...
Context 2
... symbols are detected using a 2-step process which consists of a deep learning pipeline for symbol localization followed by fine-grained recognition. As evident in Figure 3, most of the complex symbols have very similar shape, thus it is preferable to create a common class for all such symbols for symbol localization. For this, we have trained an FCN [11] based semantic segmentation model which is used to localize all such symbols. ...
Context 3
... consists of 500 annotated P&ID sheets with a 4:1 traintest ratio and is made publicly available for the benefit of research community. It includes 32 different symbols, as given in Figure 3, uniformly plotted over different graph structures which have been generated to mimic real world P&ID sheets as we have introduced different types of noise such as pixelation, blurring, salt and pepper noise in the generated sheets. The labels are assigned to symbols and pipelines maintaining the standards followed for real world P&ID. ...
Similar publications
At present usage of computational intelligence became the ultimate need of the heavy engineering industries. Digitization can be achieved in these sectors by scanning the hard copy images. When older documents are digitized are not of very high fidelity and therefore the accuracy, reliability of the estimates of components such as equipment and mat...
