樹(shù)脂鏡片缺陷瑕疵圖像獲取投影照明方案的仿真研究

15400字 33頁(yè)

摘要 近幾年中國(guó)的眼鏡產(chǎn)業(yè)發(fā)展非常迅猛，目前已躋身在世界眼鏡產(chǎn)業(yè)大國(guó)的行列中。目前中國(guó)的眼鏡生產(chǎn)企業(yè)數(shù)量較為可觀，數(shù)據(jù)表明已超過(guò)4000余家，在江蘇丹陽(yáng)以眼鏡產(chǎn)業(yè)文明全國(guó)甚至是全世界。目前生產(chǎn)的數(shù)量是可觀的，但是質(zhì)量如何呢？在鏡片生產(chǎn)過(guò)程中不可避免地出現(xiàn)各種各樣的質(zhì)量問(wèn)題。鏡片缺陷瑕疵檢測(cè)對(duì)于缺陷瑕疵的種類(lèi)、數(shù)量、大小及位置等信息的采集工作一直都是工作重點(diǎn)。目前，我國(guó)所有鏡片生產(chǎn)商的鏡片生產(chǎn)線中缺陷檢測(cè)部分全部都是由人工檢測(cè)完成的，但在很多情況下，即使對(duì)于訓(xùn)練有素的檢測(cè)人員往往也是乏味的并且是非常困難的。此外，客戶需求是在要生產(chǎn)高質(zhì)量、低成本的基礎(chǔ)上需要有更精確的缺陷檢測(cè)。這將導(dǎo)致自動(dòng)表面檢測(cè)系統(tǒng)的需求不斷增加，計(jì)算機(jī)視覺(jué)系統(tǒng)在這個(gè)不斷發(fā)展的領(lǐng)域中扮演著至關(guān)重要的角色。
樹(shù)脂鏡片缺陷瑕疵自動(dòng)化檢測(cè)設(shè)備在鏡片生產(chǎn)過(guò)程中的作用是非常重要的，它不僅能夠在生產(chǎn)鏡片的過(guò)程中提高質(zhì)量和生產(chǎn)自動(dòng)化程度，而且在鏡片生產(chǎn)過(guò)程中的能源節(jié)省方面也做出了巨大貢獻(xiàn)。鏡片缺陷自動(dòng)化檢測(cè)設(shè)備具有嚴(yán)格統(tǒng)一的鏡片檢測(cè)標(biāo)準(zhǔn)，可以有效的避免因人工識(shí)別鏡片缺陷帶來(lái)的不穩(wěn)定性的弊端，從而提高了鏡片檢測(cè)的效率和準(zhǔn)確性。而在計(jì)算機(jī)視覺(jué)應(yīng)用中，提供正確的和高品質(zhì)的照明絕對(duì)是必須的。選擇正確的照明方式仍然是一個(gè)有待攻克的堡壘，因?yàn)闆](méi)有具體的指引方法將其集成到機(jī)器視覺(jué)應(yīng)用中。事實(shí)上，任何針對(duì)最佳照明設(shè)備的選擇上的工作將增加系統(tǒng)的檢測(cè)性能，并且也將降低軟件的復(fù)雜程度。本文的主要目的便是研究鏡片圖像采集系統(tǒng)中的光照方式，開(kāi)發(fā)一種新的圖像采集系統(tǒng)，對(duì)各種鏡片缺陷進(jìn)行圖像采集，對(duì)采集的缺陷圖像進(jìn)行分析處理，最終得到具有斑、傷、羽毛、水印或雜質(zhì)等缺陷的鏡片圖像。
本文中對(duì)國(guó)內(nèi)外現(xiàn)有的照明方案進(jìn)行分析、總結(jié)并歸納，發(fā)現(xiàn)現(xiàn)如今檢測(cè)系統(tǒng)中的照明方式主要分為三種：前照光、背照光和同軸光照。此外還有散射照明和特殊照明等方式。
本文需要采用光學(xué)仿真軟件Light Tools對(duì)樹(shù)脂鏡片缺陷瑕疵圖像采集系統(tǒng)方案的點(diǎn)光源照明部分進(jìn)行仿真研究，建立光源投影照明系統(tǒng)模型和鏡片缺陷模型，獲取并分析照度圖譜，對(duì)照明系統(tǒng)中的水印缺陷的照度分布采用網(wǎng)孔優(yōu)化，驗(yàn)證實(shí)驗(yàn)結(jié)果的準(zhǔn)確性，最終獲取投影照明方案缺陷圖像，并加以分析。
關(guān)鍵詞：缺陷檢測(cè) 光照方式 軟件仿真 Light Tools

SIMULATION OF PROJECTION LIGHTING SCHEME IN IMAGE ACQUISITION SYSTEM FOR RESIN LENSDEFECT DETECTION
ABSTRACT In recent years, the development of China's optical industry is very fast, and has been among the ranks of the world's great powers in the optical industry. Present, spectacles manufacturer our country have over 4,000, the formation of a major base in Donguan, AMoy, Wenzhou, Danyang, Shanghai and Peking. Varieties of defects are inevitably in the process of lens production, type, number, size and location of information detection has been the emphasis and difficulty in the lens detection work. At present, China's lens production line defect detection part of all lens manufacturers are all done by hand detection, but, in many cases, the surface inspection tends to be tedious or difficult even for a trained operator. Moreover, the customer demands are well founded on the high no-quality costs in manufacturing and require a more precise defect detection. This causes an increasing demand for automatic surface inspection systems and the computer vision plays a crucial role in this growing field.
Lens automated defect detection equipment plays an important role in lens production process, it can not only improve the quality and degree of automation of the lens, and the lens in terms of energy savings in the production process also contributed. Automated test equipment having a lens defect lenses strictly uniform testing standards, can effectively prevent defects caused by manual identification of the instability of the lens, thereby improving the efficiency and accuracy of detection of the lens. In computer vision applications, the provision of correct and high-quality illumination is absolutely decisive. Choosing the right lighting mode is still a fortress to be overcome because there is no specific guidance on methods to integrate it into machine vision applications. In fact, any effort invested in the selection of the optimum illumination device will increase the system’s inspection performance and will also decrease the complexity of the software. The main purpose of this paper is to study the way light lens image acquisition system, to develop a new image acquisition system defect of various lenses for image acquisition, image acquisition defect processing, and end up with spots, wounds, feathers, watermarks or impurities image defects such as lenses.
Reference herein to existing lighting solutions at home and abroad for analysis, summarized and screening, is now found a total of lighting detection system is divided into: Before illumination, backlighting light coaxial illumination three. There are also scattered and special lighting, etc..
This paper uses simulation software Light Tools optical resin lenses intended for point source lighting solutions flaw defect image acquisition system simulation studies, the establishment of light projection illumination system model and defect model, obtain and analyze patterns of illumination..