Image-based Calligraphy Information Extraction And Robot Copying Research

The art of Chinese calligraphy has a long history and is profound and profound.It is a precious historical relic of the Chinese nation.It contains the culture of the Chinese nation for five thousand years and is one of the most worthy of inheritance.In modern times,due to various reasons,the calligraphy teacher is scarce,which has a great impact on the inheritance of Chinese calligraphy art.In order to promote calligraphy art and inheritance calligraphy,this paper makes the following research on the structure extraction of calligraphy characters based on the copy robot:1)The pretreatment of calligraphy characters was studied.The factors such as flying white and splashing ink in calligraphy characters were eliminated by median filtering.The holes appearing in the strokes were determined by morphological multi-factor discrimination method.The word "mouth" is distinguished and filled,which facilitates the extraction of strokes and skeletons in the later stage.2)Research on the stroke extraction of Chinese characters.In order to obtain the writing information of strokes,this paper starts with the Chinese character skeleton and uses the parallel refinement algorithm to extract the basic skeleton.This kind of skeleton has the problems of incomplete refinement and distortion at the intersection.This paper uses the construction template to match,in the skeleton.The redundant points are eliminated to obtain a fully refined skeleton.On the basis of obtaining the completely refined skeleton,the distorted parts are deleted,and then the stroke segments belonging to the same stroke are extracted and interpolated and combined by the method of local correlation degree to extract the single stroke skeleton.3)The Chinese character width information and the stroke order information are extracted.For the width information,the maximum circle method is used to extract,with the skeleton point as the center,and the maximum circle diameter within the edge of the Chinese character is regarded as the pen width.For the extraction of stroke order information,this paper constructs a Chinese character structure segmentation method based on connected domain,which divides the Chinese character into individual structural components,and specifies the writing order of each component by the distance between the centroid of the structural component and the origin.Similarly,through The center of the stroke is far from the origin to determine the writing order for the strokes in each part;finally,on the basis of the writing rules of “from top to bottom” in most strokes,combined with “horizontal” and “lifting”,etc."Special stroke recognition of writing,determine the starting and ending point of each stroke.4)Finally,under the experimental platform of 3D Cartesian robot,the mathematical model of writing depth and pen width information of brush is established by experimental method.Combining with the previous work,the experiment of robot calligraphy is completed.