| 摘要: | 以目前工廠自動化的發展情況來說，機器視覺技術的使用算是一項重要的發展主力。機器視覺應用在工業上的例子很多，例如產品尺寸檢測、產品包裝作業、LED亮點檢測、機械手臂焊接作業及產品二維條碼之辨識。以上的例子若是以傳統的方式來完成，機器視覺則佔有準確性高、生產速度快、長時間作業、作業安全性、建構及維修成本低及學習速度快等的優勢。事實上其於工業自動化方面的應用並不止於此，所以本論文就取其中一項來進行研究，號碼辨識在學術研究上的辨識方法有很多。本研究之目的在探討利用影像面積統計法及影像投影法此兩項方法，去提升號碼辯識的辯識率及提供號碼辯碼方法。另一方面，我們利用自動化設備定位精準及穩定性程度高的優點，計算工件移動速度於正確的時間及位置取得影像，經由程式讀取及前置處理Sobel濾波清除影像noise後，使用線掃描方法計算統計出字型面積及字型特徵，再正確分辯出影像字型。我們為了讓本研究為更符合實際作業情況，乃採用已使用於工廠的自動化打刻機、工業字體之鋼模，配合CCD影像擷取系統模擬工廠作業環境，以達到實際的結果，期望達到能應用於工廠自動化的目的，也希望藉由本研究提供與一般論文在號碼辨識方面另一種不同的思考方向。 As to the development of automation of the factory at present, the use of the vision technology of the machine can be regarded as an important development main strain. Machine Vision has been applied to industrial in many conditions, for example measurement measuring of the products , the homework of packing of product , LED light spot are measured , the robot welds work and two-dimentional bar code of products. If the above-mentioned examples come by way of tradition and finish, the vision of the machine occupies accuracy high, work security for a long time, building constructing and maintaining the advantage with low costs and studying the speed fast etc. with quick speed of production. In fact its application in the industrial automation is combined not only herein, so a thesis fetches one of them to carry on research, there are a lot of methods of distinguishing that the number distinguishes on the academic research. The purpose of this research is to probe into and utilize the statistic law of the area of image and this two methods of projection law of the image , the ones that raised numbers to debate and know debates the knowing rate and offers numbers to debate one yard of methods . On the other hand, we utilize automation equipment to orient the advantage of perfect and stability with high degree, calculate that the work piece moves the speed and makes the image in correct time and position, via the procedure reads and deal with Sobel leadingly after straining waves and removing image noise, use the line to scan the method and calculate and count the area of style of calligraphy and characteristic of style of calligraphy, and then explain for out the style of calligraphy of the image correctly. For let research this for correspond to reality homework situation even more , is it use automation in factory is it carve plane , industrial steel mould of script to play already to adopt we, cooperate with CCD image to pick and fetch the systematic homework environment of simulation factory, in order to reach the real result, expect to achieve the purpose that can be applied to automation of the factory, hope to make use of and offer by this research and distinguish another kind of different thinking directions of the respect with the general thesis in the number too. |
