论文部分内容阅读
针对机器人视觉系统对抓取物的模糊识别问题,参照人眼-脑识别对象的过程,建立了包括区域分割与模糊识别两个环节的识别算法.以视觉图像的灰度与色度特征作为区域分割与模糊识别的依据,从灰度、色度及形体特征上提取特征集的指标,并根据经walsh变换后图像灰度迅速向低频聚集的特点,提出基于walsh变换的基元模式识别特征的定义方法.在构建识别目标矩阵与关系矩阵的基础上,应用模糊关系合成与最大隶属度原则建立识别算法.该算法可从少量的采样点中识别出对象,具有较好的实时性.
Aiming at the fuzzy recognition of the grasping robot vision system and referring to the human-brain recognition process, the recognition algorithm including the segmentation and the fuzzy recognition is established. Taking the grayscale and chroma features of the visual image as the region Segmentation and fuzzy recognition, the features of feature set are extracted from the grayscale, chromaticity and shape features, and based on the characteristics of walsh-transformed image grayscale quickly converging to low frequency, the feature of primitive recognition based on walsh transform Based on the construction of recognition target matrix and relation matrix, the recognition algorithm is established by applying the principle of fuzzy relation composition and maximum membership, which can identify objects from a small number of sampling points and has better real-time performance.