【摘要】：隨著機(jī)電一體化的發(fā)展,零件大批量的生產(chǎn)要求不斷提高。高速、高精度不斷滲透到機(jī)加工的各個生產(chǎn)環(huán)節(jié),然而在加工過程中經(jīng)常由于外界及材料本身特性的影響,導(dǎo)致生產(chǎn)出來的銷釘存在表面劃傷、銹跡、尺寸超出公差范圍等缺陷,將影響銷釘?shù)恼Ｊ褂?對于體積較小的銷釘,人工對合格零件分揀已經(jīng)不能滿足工業(yè)生產(chǎn)對產(chǎn)品質(zhì)量以及生產(chǎn)速度的要求。因此,基于圖像處理對銷釘進(jìn)行非接觸檢測篩選的技術(shù)研究非常重要。通過對眾多邊緣算法的研究,對銷釘邊緣精準(zhǔn)定位的算法選擇中,理論的插值邊緣檢測算法,在實(shí)際圖像檢測過程中,存在定位不準(zhǔn)確的缺點(diǎn)。本文針對高斯亞像素算法無法適用于實(shí)際檢測情況的缺點(diǎn),在對銷釘?shù)暮Y選過程中加入了邊緣檢測步驟,對原始的插值算法進(jìn)行了優(yōu)化。本論文基于MATLAB編程平臺,將算法實(shí)現(xiàn)。并對各種經(jīng)典的邊緣算子進(jìn)行比較,在銷釘表面質(zhì)量和邊緣定位過程中,確定對SSIM結(jié)構(gòu)相似度函數(shù)和高斯亞像素插值算法進(jìn)行優(yōu)化,并以對應(yīng)的程序?qū)⑺惴▽?shí)現(xiàn)。處理過程如下:首先通過Canny算法對圖像進(jìn)行處理,得到圖像邊緣輪廓,將該輪廓與標(biāo)準(zhǔn)銷釘?shù)妮喞M(jìn)行相似度判斷,完成初步的篩選;然后,在針對銷釘尺寸測量過程中對邊緣的角度進(jìn)行旋轉(zhuǎn),對變換后的圖像進(jìn)行高斯插值計算,精確的針對銷釘?shù)某叽邕M(jìn)行篩選。這兩種邊緣檢測算法的結(jié)合,在對邊緣檢測過程中具有方向不變性,而且能夠高效率的對銷釘進(jìn)行篩選。實(shí)驗(yàn)證明將該算法運(yùn)用于銷釘邊緣提取能滿足精度要求。
[Abstract]:With the development of mechatronics, the requirement of mass production of parts is increasing. High speed and high precision continuously permeate into every production link of machining. However, in the process of processing, the pin produced often has surface scratches and rusts due to the influence of the outside world and the characteristics of the material itself. Such defects as dimension exceeding tolerance range will affect the normal use of pins. For smaller pins, manual sorting of qualified parts can no longer meet the requirements of industrial production for product quality and production speed. Therefore, it is very important to study the non-contact detection and screening of pins based on image processing. Through the study of many edge algorithms, the theoretical interpolation edge detection algorithm in the selection of pin edge accurate location algorithm, in the actual image detection process, there are shortcomings of inaccurate location. Aiming at the disadvantage that the Gussia pixel algorithm can not be applied to the actual detection, the edge detection step is added to the selection process of pin, and the original interpolation algorithm is optimized. This paper is based on MATLAB programming platform, the algorithm is implemented. In the process of pin surface quality and edge location, the similarity function of SSIM structure and Gauss pixel interpolation algorithm are optimized, and the algorithm is implemented by corresponding program. The processing process is as follows: firstly, the image edge contour is obtained by Canny algorithm, and the similarity between the contour and the standard pin contour is judged, and the preliminary screening is completed. Then, the angle of the edge is rotated in the process of pin size measurement, the transformed image is calculated by Gao Si interpolation, and the size of pin is screened accurately. The combination of these two edge detection algorithms is directionally invariant in the process of edge detection and can screen pins efficiently. The experimental results show that the algorithm can meet the precision requirement when it is applied to pin edge detection.
【學(xué)位授予單位】：天津工業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 許光明;吳昭;周春蘭;沈軍民;;基于視覺的零件尺寸測量[J];工業(yè)控制計算機(jī);2015年04期

2 劉明周;馬靖;張淼;趙志彪;楊鐸;王強(qiáng);;基于機(jī)器視覺的機(jī)械產(chǎn)品裝配系統(tǒng)在線作業(yè)方法[J];計算機(jī)集成制造系統(tǒng);2015年09期

3 呂繼武;于偉;鄭偉;;一種基于改進(jìn)的HOUGH變換的車輪檢測方法[J];電子技術(shù)與軟件工程;2014年13期

4 崔然;;淺析計算機(jī)視覺技術(shù)在農(nóng)產(chǎn)品檢測及分級中的應(yīng)用[J];電子測試;2013年09期

5 祁曉玲;趙霞霞;靳伍銀;;基于機(jī)器視覺的軸類零件幾何尺寸測量[J];組合機(jī)床與自動化加工技術(shù);2013年01期

6 來躍深;陳琛;田軍委;程鋼;;高斯插值亞像素邊緣檢測算法的優(yōu)化[J];西安工業(yè)大學(xué)學(xué)報;2012年10期

7 林雯;;基于計算機(jī)視覺和神經(jīng)網(wǎng)絡(luò)的芒果外觀等級分類研究[J];安徽農(nóng)業(yè)科學(xué);2010年23期

8 高晶晶;李景茹;劉瑞敏;;計算機(jī)視覺技術(shù)在花生破損檢測中的應(yīng)用[J];機(jī)電產(chǎn)品開發(fā)與創(chuàng)新;2010年03期

9 費(fèi)麗君;譚峰;;機(jī)器視覺技術(shù)在大豆葉片葉綠素含量測算上的應(yīng)用[J];農(nóng)機(jī)化研究;2010年03期

10 凌遠(yuǎn)煥;徐杜;蔣永平;劉長紅;黃杰賢;;基于局部區(qū)域灰度矩圖像邊緣定位方法的研究[J];光學(xué)與光電技術(shù);2009年03期

，

本文編號：2406773