机器人视觉目标数字图像实时处理及分割

The Real-Time Disposal and Segmentation of Robot Vision Target Digital Image

机器人视觉目标图像信噪比低、背景噪声干扰大,目标识别处理通常利用目标的灰度信息进行预处理。文中设计一种基于数学形态学和遗传算法的灰度图像实时预处理和阈值处理技术图像分割方法。描述图像的基本结构和特征,用具有一定形态的结构元去度量和提取图像中的对应形态,以达到对图像分析和识别;经过预处理的原始图像,采用基于遗传算法的最大类间方差图像分割法,非线性快速地查找到最优的分割阈值,从噪声图中分割出可能目标。基于ARM嵌入式微处理器,以复杂可编程逻辑器件CPLD作为时序控制单元,进行图像处理、分割,实现了图像信息的采集、存储、传输及处理为一体。仿真实验表明,该方法实时性好,简捷、快速,对运动目标的图像识别有较好的实用价值。

The robot vision target image has low SNR （ Signal-to Noise） and high background noise disturbance, therefore the target recognition disposal must make use of the target＇s grey information. This paper designs the method of grey chroma on real-time and threshold value disposal image segmentation based on mathematical morphologic and genetic arithmetic. It can describe basic structure and characters of the image, using some structural units which have some conformation to measure and distill the corresponding conformation. The square margin dynamic threshold value segmentation method is used to deal with the original image which was pre-disposed, looking for the best segmentation threshold value non-linearly and fast, then the possible object from the noise images is divided. Based on the ARM embedded micro-processor, using the programmable logical device CPLD as scheduling controller unit to conduct image disposal, segmentation and matching program, this method integrates with the image acquisition, storage, transmission and disposal. The emulate experiment shows that this method is brief and fast. It has good real time response and great practical value to recognize the image of moving object.