Object-process based segmentation and recognition of ANSI and ISO standard dimensioning texts

Abstract

Recognition of dimensioning text in engineering drawings is an essential part of the dimension understanding process, since this text is an important component of the dimension-set. We first introduce the OPD expression of the structural relations among the dimension-set components and the illustration of ISO and ANSI drafting standards. We then present principles and implementation of a method of segmentation and recognition of dimension text in ANSI and ISO based drawings. The method is vector-based, implying that the input is a set of vectors-wires (bars and arcs)-resulting from the orthogonal zig-zag vectorization, arc segmentation, and arrowhead pair recognition. Initial textbox extraction is done by a region growing process, performed on text-wire candidates. On the basis of textbox context (neighboring annotation wires), the drafting standard is detected. Raw textboxes are divided into logical textboxes, which are further decomposed into basic textboxes. A neural network based OCR algorithm is applied to each single character image decomposed from the basic textboxes and the symbol in each basic textbox in recognized. Finally, the OCR recognition results are verified by using contextual information and comparing the results with the measurements made directly on the drawing.