【正文】
本文所設(shè)計(jì)的圖書館機(jī)器人不要求為每本書都配備一個(gè)圖書存儲(chǔ)盒,而是充分利用視覺識(shí)別優(yōu)勢(shì),并巧妙設(shè)計(jì)RFID電子標(biāo)簽數(shù)據(jù)格式,提出一種由模糊到精確定位的機(jī)械手定位方法,采用多源信息融合法提高定位的魯棒性,實(shí)驗(yàn)結(jié)果表明,機(jī)械手可實(shí)現(xiàn)精確定位與快速抓取。本系統(tǒng)采用Canny算子做書脊邊緣檢測(cè),采用累計(jì)概率霍夫變換(PPHT)算法實(shí)現(xiàn)對(duì)書脊線段的查找,用VC++和OpenCV開發(fā)了視覺檢測(cè)與識(shí)別程序,經(jīng)多次實(shí)驗(yàn),反復(fù)調(diào)整閾值,得出的比較理想的檢測(cè)效果如圖5所示: (a)原始圖像 (b)視覺書脊檢測(cè)結(jié)果圖5 Canny PPHT法書脊檢測(cè)效果圖 經(jīng)程序處理,可去除干擾線段,分離出有效書脊線段,但從檢測(cè)效果看,魯棒性仍不是很強(qiáng),因此想到與機(jī)械手讀取到的RFID射頻信號(hào)“書架層內(nèi)區(qū)間相對(duì)位置號(hào)”相融合,增強(qiáng)識(shí)別的魯棒性。本文提出一種基于視覺的分區(qū)搜索法。圖像坐標(biāo)系和世界坐標(biāo)系之間的關(guān)系為: ZcBuBvB1 = αusu00αvv0001RtXwBYwBZwB1 = KRtXwBYwBZwB1 = PXwBYwBZwB1 (式2)式2中,f為攝像機(jī)的焦距,αu= f/dx, αv= f/dy,s = s39。(3) 圖書入庫(kù)圖書管理信息系統(tǒng)通知書架確定空閑位置給待入庫(kù)圖書,引導(dǎo)機(jī)器人將圖書放到指定位置。圖書館內(nèi)道路按照館藏類型劃分成多個(gè)區(qū)域,每個(gè)區(qū)域有相應(yīng)的路面標(biāo)志,車載無線圖像采集模塊獲得路面標(biāo)志圖像并通過無線網(wǎng)絡(luò)傳送給上位機(jī),上位機(jī)程序進(jìn)行圖像處理,并計(jì)算AGV控制參數(shù),再通過無線網(wǎng)絡(luò)發(fā)送至機(jī)器人主控制器調(diào)速系統(tǒng),實(shí)現(xiàn)PWM直流電機(jī)調(diào)速控制[46]。關(guān)鍵詞:圖書館機(jī)器人,機(jī)械手,物聯(lián)網(wǎng),機(jī)器視覺,RFID文獻(xiàn)標(biāo)識(shí)碼:A 中圖分類號(hào):TP273A Research on the Accurate Position Method for the Library Robot Manipulator based on Machine VisionMingfang Du1,2, Jianjun Fang1, Junzheng Wang2 ( of Automation, Beijing Union University, Beijing 100101 of Automation, Beijing Institute of Technology, Beijing 100081)【Abstract】Firstly the cooperative work process and the automation work flow of the no man keeping watch library taking the robot as its central and using the internet of things technology are introduced in this paper. A new data format of the RFID electronic label for books is designed, and the robustness of positioning of the manipulator is improved by the information fusion of the RFID label source and the CCD sensor source installed on the manipu