When acquiring symbol-coded structured-light images, the shape of the tested object in the image is formed by masses of symbols. In order to obtain 3D information about the tested object, each symbol should be recognized and to obtain detailed symbol information, a series of image proceedings are needed. When using the threshold process image, the effect of one part will be influenced by the other parts. Therefore the image segmentation is poor. Typical segmentation methods commonly operate in a single connected domain. In order to segment a shape which is composed of masses of domains, a new method based on mathematical statistics has been developed in this work. The method can automatically extract the main content of an image by a contour. The contour size and shape are controlled by two parameters. Through experiment, the method divides the image into two regions and processes the foreground region. The effect of this method is better than direct image processing and it enhances the symbol recognition ratio.
WANG Xiao;ZHANG AiJun.
Area segmentation for coded-light acquired pictures[J]. Journal of Beijing University of Chemical Technology, 2015, 42(1): 107-112
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]马玉坤, 王中亚, 杨国威, 等. 基于线结构光传感器的公路平整度测量系统[J]. 传感技术学报, 2013, 26(11): 1597-1603. Ma Y K, Wang Z Y, Yang G W, et al. A system based on structured-light sensors for measurement of pavement evenness[J]. Chinese Journal of Sensors and Actuators, 2013, 26(11): 1597-1603. (in Chinese) [2]Albitar C, Graebling P, Doignon C. Robust structured light coding for 3D reconstruction[C]∥International Conference on Computer Vision, Rio de Janeiro, Brazil, 2007: 1-6. [3]彭祎帆, 陶毅阳, 于超, 等. 基于红外结构光的三维显示用交互装置[J]. 光学学报, 2013, 33(4): 0412005-1-0412005-7. Peng Y F, Tao Y Y, Yu C, et al. Three-dimenaional display interaction device based on infrared structured light[J]. Acta Optica Sinica, 2013, 33(4): 0412005-1-0412005-7. (in Chinese) [4]Otsu N. A threshold selection method from gray-level histograms[J]. IEEE Transections on Systems, Man and Cybernetics, 1979, 9(1): 62-66. [5]Farid H, Simoncelli E P. Optimally rotation-equivariant directional derivative kernels[C]∥International Conference on Computer Analysis of Images and Patterns, Keil, Germany, 1997: 207-214. [6]Shapiro L G, Sockman G C. Computer vision[M]. Upper Saddle River, USA: Prentice Hall, 2001: 306-310. [7]Shi J B, Malik J. Normalized cuts and image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 888-905. [8]Kass M, Witkin A, Terzopoulos D. Snakes: active contour models[J]. International Journal of Computer Vision, 1988, 1(4): 321-331. [9]纳跃跃, 于剑. 一种用于谱聚类图像分割的像素相似度计算方法[J]. 南京大学学报: 自然科学版, 2013, 49(2): 159-168. Na Y Y, Yu J. A new pixel affinity for sectral image segmentation[J]. Journal of Nanjing University: Natural Sciences, 2013, 49(2): 159-168. (in Chinese) [10]倪麟, 龚劬, 曹莉, 等. 基于自适应加权中值滤波的二维Otsu图像分割算法[J]. 计算机应用研究, 2013, 30(2): 598-600. Ni L, Gong Q, Cao L, et al. Two-dimensional Otsu image segmentation algorithm based on adaptive weigted median filter[J]. Application Research of Computers, 2013, 30(2): 598-600. (in Chinese) [11]Ram S, Rodríguez J J. Random walker watersheds: a new image segmentation approach[C]∥International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vancouver, Canada, 2013: 1473-1477. [12]Zeng L, Chen J, Li Z G, et al. 3D image spectral segmentation based on the region mean histogram[C]∥International Congress on Image and Signal Processing, Hangzhou, 2013: 527-531. [13]Yu Q C, Feng H M, Zhang H X. Multi-resolution decoding method of symbol M array surface structured light[C]∥International Conference on Artificial Intelligence and Computational Intelligence, Sanya, Hainan, 2010: 63-68. [14]李想. 3种照度计算方法的比较[J]. 上海电力学院学报, 2011, 27(1): 83-86. Li X. Comparison of three lumination calculation methods[J]. Journal of Shanghai University of Electric Power, 2011, 27(1): 83-86.