【作者】 刘业胜；

【导师】 陈明君；

【作者基本信息】 哈尔滨工业大学 ， 机械制造及其自动化， 2010， 硕士

【摘要】 随着科学技术的发展,光学成像系统在国防尖端装备及民用工业中应用范围越来越广,人们对其要求也不断提高。如智能武器、智能机器人视觉系统以及微型飞行器等,期望整个系统重量轻、体积小、视场大以及对运动目标更敏感。近年来,由于系统的机械装置和电子器件进一步小型化,轻量化已变得越来越困难。因此许多科研机构想通过缩小光学系统部位从而使整个系统达到轻量化与小型化。仿生光学复眼具有体积小、重量轻、视场大、灵敏度高、可测速等优点,因而引起国内外学者的广泛关注。各国学者纷纷对光学复眼展开研究,并取得了一定的研究成果,但其成像仍然存在一定的缺点。因此,怎样进一步提高光学复眼的成像特性变得尤为重要,给仿生光学复眼的研究提出了新的挑战。本文首先通过对仿生光学复眼进行光学理论分析,得到了仿生光学复眼小眼的三维空间函数关系式。该仿生复眼小眼采用高次拟合曲面取代球面阵列小眼,提高小眼在近光轴以外部分的成像特性,克服了球面小眼成像过程中中间清晰边沿模糊的缺点,改善了复眼的成像效果。其次,通过小眼三维空间函数生成的点云,利用Pro/E逆向工程实现了小眼的初始网格化。采用基于三角形递归分割插补算法的RecurSurface(重构曲面)系统完成了小眼的重构,重构精度达到2nm。通过RecurSurface(重构曲面)系统与Pro/E的结合,实现了整个复眼的造型。该仿生光学复眼整体尺寸为19×17×3mm,由七个近似六边形的小眼组成,以中间小眼为基准,其余六个小眼分别绕中间小眼六条边旋转20°。近似六边形的小眼组合,消除了球面小眼组合过程中存在的间隙,克服了球形小眼复眼成像过程中存在“麻点”的不足,进一步改善了复眼的成像特性。最后,研究了加工路径及加工参数对仿生光学复眼加工精度的影响,并应用Pro/E数控仿真模块对构建的仿生光学复眼模型进行数控仿真及代码后置处理,生成了可用于实际加工的复眼G-代码。同时,为了消除实际加工中可能出现的错误,利用Vericut对Pro/E仿真生成的G-代码进行验证。将G-代码进行格式变化,生成PMAC控制器能够识别的数控代码,利用基于PMAC控制的三轴微细机床进行加工,得到仿生光学复眼模型并对加工误差进行分析。更多还原

【Abstract】 With the development of science and technology,optical imaging system has been widely and increasingly used in advanced equipments of national defence and civilian industries, and much higher requirement is needed. Such as smart weapons, intelligent robot vision system and micro air vehicles, expected the whole system weight light, size small, large field of view and the more sensitive to the moving target. However, in recent years the system of mechanical devices and electronic devices has become increasingly difficult to be further miniaturization and much lighter. So many scientists want to reduce the optical system parts in order to make the whole system the lighter weight and further miniaturization. And with the advantages of small volume, light weight, large field-of-view, high sensitivity and measuring speed, the artificial optical compound eye has got more and more attention. The optical compound eye has been researched and lots of results have been made. But the image of the eyes still has some shortcomings. Therefore, how to make further improvement of the optical compound eye imaging properties is becoming particularly important and new challenges to study the compound eyes have been raised.In this paper, firstly, the three-dimensional function of the compound eye’s ommatidium can be got through the analysis of artificial optical compound eye by optical theory. The ommatidium is described by higher-order curve, which replace with the high fitting spherical surface array and the imaging features of the outside optical axis part is enhanced. By that way, it overcome the disadvantage that the image is clear in middle but fuzzy in edge with the sphere array and improve the image effect of compound eye.Secondly, by the three-dimensional function of the ommatidium, the point cloud is calculate. The initial grid of ommatidium is generated through Pro / E Reverse engineering. Then the ommatidium model is reconstructed by RecurSurface (Surface Reconstruction) system which is base on Triangle recursive subdivision algorithm and the reconstruction accuracy can catch up with 2nm. The model of the compound eye is reconstructed through the RecurSurface (Surface Reconstruction) system and Pro/E. The overall size of the optical compound eye is 19×17×3mm. The eye is formed by seven ommatidium, and base on the middle one, others rotate 20°around its sides. The eye is assembled by approximate hexagonal ommatidiums, which eliminate the gaps that produced by combination of the spherical ommatidiums, and it overcome the image shortage of“Ma points”which appear when the ommatidiums is sphere. In this way, the imaging characteristics of the compound eye can be further improved.Finally, the processing path and processing parameters are be studied which are very important to processing accuracy of the artificial optical compound eye. And then the NC manufacturing simulation and post processor of the eye model after subdivision are completed adapting the NC machining modular of Pro/E. The G-code of eye is generated. Meanwhile, in order to eliminate the possible errors in the actual processing, the G-code is checked with Vericut soft. After the G-code format is changed to the NC code that the generating PMAC controllers can identify, the eye model is machined by PMAC-based three-axis control machine tool. At last the processing errors are analyzed.更多还原