一种物料袋场景图像增强算法.docVIP

一种物料袋场景图像自适应增强算法( 贾 伟1，刘彦宾1，刘 炜1,2，董新久1,杨茂强1 （1.遵义师范学院计算机与信息科学学院, 贵州 遵义 563002；2.贵州大学 计算机科学与技术学院,贵阳563003） 摘要：为了提高自然场景中物料袋的分割、检测与识别效率，提出了一种基于模糊理论的物料袋场景图像自适应增强算法。该算法首先运用otsu算子对图像背景与目标进行自适应分类，并确定目标与背景图像像素的最佳渡越点；然后采用新的线性隶属度函数将图像对比度值从灰度域变换为模糊域，通过渡越点计算广义增强算子(GFO)与双曲正切函数的关联参数，并利用广义增强算子与双曲正切函数对图像进行增强变换；最后采用线性变换与灰度值叠加的方法将变换值由模糊域映射为灰度域。实验结果显示，该算法在减少迭代次数的同时，使图像局部细节得到有效增强，在实现算法自适应的条件下，可以得到良好的视觉效果。 关键词：物料袋；图像增强；模糊集理论；双曲正切函数 中图分类号：TP391.41 Adaptive enhancement algorithm of the image of materials bag scene( JIA Wei1, LIU Yan-bin, LIU Wei, DONGXin-jiu,YANG Mao-qiang (1.Department of Computer and Information Science,ZunYi Normal College,563002,Zun Yi;2.College of Computer Sience Technology, Gui Zhou University, 563003 Gui Yang) Abstract: In order to improve segmentation,?detection and recognition efficiency of the materials?image bag?, the paper proposed an adaptive enhancement algorithm of materials bag scene based on fuzzy theory. Firstly, it classifies the image?background and objectives by using the otsu?operator, then determines the crossover point?of the objectives and the background?image. Secondly, a novel linear membership function is applied to map the image gray value into fuzzy?domain, then computes the correlation parameter of the generalized?enhancement operator?(GFO)?and the hyperbolic tangent function by the crossover point?, the enhancement result is?transformed by using the generalized?enhancement operator?and?the hyperbolic tangent function. Finally, it uses the method of superposition of the linear?to map the fuzzy values ?into ?the gray domain. The experimental result shows that,?the?local image?details?is effectively enhanced while the algorithm uses fewer iterations, under the realization of?algorithm of adaptive?conditions,?the algorithm also has better visual effect. Key words：materials?bag; image enhancement; fuzzy set theory; hyperbolic tangent function.0 引言 视觉技术的发展已从实验室逐渐走向实际应用[1-3]。特别在物流领域，随着该领域的自动化、智能化水平的提高，机器人越来越多的被用来处理复杂多变的工程模型以及形状