Abstract: Tire prints and wear characteristics segmentation is the prerequisite of the computer-aided tire prints recognition system which can confirm vehicles in an accident. According to the characteristics of tire image acquired at the scene, a kind of combination of grey theory and particle swarm algorithm of normalized image segmentation is put forward. This method uses the grey relation analysis to measure the similarity between pixels, and derives from normalized criteria to the dual-threshold criteria. Experimental results show that the algorithm can have an effective segmentation to tire prints image and underline the tire wear features, especially have a high calculating speed, and lay the foundation for the next feature extraction and recognition.

Key words: tire prints, grey theory, particle swarm algorithm, normalized criteria

摘要： 轮胎印痕磨损特征分割是计算机辅助轮胎印痕识别系统确定事故车辆的前提条件。根据现场采集的轮胎印痕图像的特点，提出了一种结合灰色理论和粒子群算法的归一化图像分割。该方法使用灰色关联分析来衡量像素点间的相似性，将归一化准则推导为双阈值分割准则。实验表明：该算法能对轮胎印痕图像进行有效的分割，突显出轮胎的磨损特征信息，计算速度快，为下一步特征提取和识别奠定了基础。

关键词: 轮胎印痕, 灰度理论, 粒子群算法, 归一化准则