栗俊艷,米月琴,弓 靖,劉鵬飛
(航天科技集團(tuán)九院第16研究所 機(jī)器人事業(yè)部,陜西 西安 710100)
一種基于機(jī)器視覺(jué)的工業(yè)機(jī)器人分揀系統(tǒng)
栗俊艷,米月琴,弓靖,劉鵬飛
(航天科技集團(tuán)九院第16研究所 機(jī)器人事業(yè)部,陜西 西安710100)
摘要針對(duì)傳統(tǒng)的人工分揀作業(yè)效率低、質(zhì)量難以保證的問(wèn)題,提出了一種基于機(jī)器視覺(jué)的工業(yè)機(jī)器人分揀系統(tǒng),該系統(tǒng)利用圖形分割和圖像匹配相結(jié)合的算法,克服了傳統(tǒng)視覺(jué)算法受光照和噪聲影響大的問(wèn)題,實(shí)驗(yàn)系統(tǒng)通過(guò)采集棋子及場(chǎng)景圖像并進(jìn)行圖像處理,提取黑白棋子的坐標(biāo)以及分類(lèi)信息,實(shí)施坐標(biāo)轉(zhuǎn)換,維護(hù)分揀棋子的運(yùn)動(dòng)蹤跡,最終控制機(jī)器人實(shí)現(xiàn)分揀棋子的動(dòng)作。實(shí)驗(yàn)結(jié)果表明,該控制系統(tǒng)軟硬件設(shè)計(jì)正確,分揀成功率高,且易于實(shí)施。
關(guān)鍵詞機(jī)器視覺(jué);工業(yè)機(jī)器人;分揀;坐標(biāo)系
Sorting Technology Based on Industrial Robot of Machine Vision
LI Junyan,MI Yueqin,GONG Jing,LIU Pengfei
(16thInstitute,China Aerospace Science and Technology Corporation,Xi’an 710100,China)
AbstractA sorting system based on industrial robot of machine vision is proposed to solve the low efficiency and poor quality control of the traditional manual sorting operation.The system combines the image segmentation with image matching to overcome the disadvantages that traditional visual algorithm is affected by light and noise.The experiment system collects chess pieces and scene image and carries on the image processing,extracts the coordinates and classification of the black and white chess pieces,implements the coordinate conversion,maintains the movement trail of the sorting chess pieces,and controls the movement of the robot to realize the sorting action.The experimental results show that the hardware and software design of the control system is correct,and the success rate of the sorting is high with easy implementation.
Keywordsmachine vision;industrial robot;sorting;coordinate system
工件分揀是工業(yè)生產(chǎn)環(huán)節(jié)重要的組成部分,其目的是將不同類(lèi)型的物料或工件分類(lèi)擺放到相應(yīng)的位置,其步驟主要分為定位、識(shí)別、抓取和放置4個(gè)階段[1]。傳統(tǒng)中一般采用人工分揀的方式,因工人操作存在疲勞度的問(wèn)題,對(duì)于長(zhǎng)時(shí)間作業(yè)來(lái)說(shuō),顯然無(wú)法滿(mǎn)足,即使采用工業(yè)機(jī)器人方式,由于傳統(tǒng)的工業(yè)機(jī)器人一般采用示教或離線(xiàn)編程的方式工作,機(jī)器人到達(dá)目標(biāo)點(diǎn)的位姿是固定的,每次動(dòng)作均為重復(fù)進(jìn)行,只能重復(fù)完成預(yù)先規(guī)定好的動(dòng)作,對(duì)于加工對(duì)象以及工作環(huán)境的感知能力低[2],使機(jī)器人的使用受到了限制。隨著機(jī)器視覺(jué)和人工智能的發(fā)展,機(jī)器人也能夠適應(yīng)分揀環(huán)境的需求,隨時(shí)變更作業(yè)對(duì)象和分揀工序。……