周全+李濤
摘 要: 針對傳統(tǒng)判斷方法一直存在判斷誤差大、效率低的問題,提出基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法。對于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性用權(quán)值進(jìn)行表示,把遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)中存在的二維屬性進(jìn)行提取,并進(jìn)行合理性的辨別,引入快速魯棒識(shí)別算法,分析遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性權(quán)值,并且通過權(quán)值對比有效地對遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性進(jìn)行判斷。實(shí)驗(yàn)結(jié)果表明,改進(jìn)判斷方法能夠有效地辨別設(shè)計(jì)的合理性,保證了數(shù)據(jù)的有效性以及計(jì)算的準(zhǔn)確度。
關(guān)鍵詞: 遠(yuǎn)程圖像識(shí)別; 景觀設(shè)計(jì)合理性; 權(quán)值對比; 判斷誤差; 魯棒識(shí)別; 二維屬性
中圖分類號: TN911.73?34; TP311 文獻(xiàn)標(biāo)識(shí)碼: A 文章編號: 1004?373X(2018)04?0154?03
Abstract: In allusion to the big judgment error and low efficiency of the traditional judging method, a landscape design rationality judgment method based on remote image recognition is proposed. The weight values are used to represent the rationality of remote image recognition based landscape design. The two?dimensional attribute existing in remote image recognition based landscape design is extracted for rationality judgment. The fast robust recognition algorithm is introduced to analyze the weight values of remote image recognition based landscape design. The weight values are contrasted to effectively judge the rationality of remote image recognition based landscape design. The experimental results show that the improved judgment method can effectively distinguish the rationality of the design so as to ensure the validity of the data and the accuracy of the calculation.
Keywords: remote image recognition; landscape design rationality; weight value contrast; judgment error; robust recognition; two?dimensional attribute
利用遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)過程中,由于使用的是遠(yuǎn)程圖像,因此會(huì)產(chǎn)生一定的視覺誤差,為此,對遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性進(jìn)行判別是十分重要的[1]。遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性包括:景觀建筑結(jié)構(gòu)的安全性和施工過程的誤差性以及設(shè)計(jì)的景觀樣式是否有足夠的技術(shù)支撐。遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性能夠影響結(jié)構(gòu)整體施工過程,同時(shí)也是人員調(diào)配以及施工準(zhǔn)備的重要數(shù)據(jù)指標(biāo)。目前對遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性的判斷方法大致包括以下幾種:比例尺權(quán)衡法、超聲波對照法、攝影圖像法等[2],但是對多幾何形體以及多遮擋物的景觀其判斷的準(zhǔn)確性非常的低。針對上述過程中遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性,本文提出一種有效的判斷方法,并通過實(shí)驗(yàn)數(shù)據(jù)有效證明驗(yàn)證了本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別景觀設(shè)計(jì)合理性判斷方法的有效性。
1 遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性加權(quán)表示
1.1 遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性提取
本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,使用權(quán)值進(jìn)行表示,但是需要進(jìn)行相關(guān)數(shù)據(jù)提取才能夠使用權(quán)值進(jìn)行有效表示,首先進(jìn)行圖像有效數(shù)據(jù)確認(rèn),過程如下:
式中:[R2YIU]為二維參考圖像的數(shù)據(jù)確認(rèn)過程;[Ai]為參考圖像的辨別值;[Bi]為參考圖像的提取值;[Ai2],[Bi2]分別為關(guān)聯(lián)圖像的處理參考值以及可用辨別值。經(jīng)過上述的圖像確認(rèn)能夠?yàn)橛行У貙D像的數(shù)據(jù)提取做好準(zhǔn)備工作[3?4]。提取一項(xiàng)數(shù)據(jù)是輪廓數(shù)據(jù),提取過程如下:
式中:[f1]為景觀局部邊緣誤差的掩蓋值;[N2]為設(shè)計(jì)的景觀邊緣輪廓特征量;[pi,j]為景觀圖像邊緣平滑灰度特征。經(jīng)過邊緣數(shù)據(jù)的提取過程,會(huì)使SVM函數(shù)對設(shè)計(jì)景觀的色差、陰影、節(jié)點(diǎn)、光感進(jìn)行數(shù)據(jù)提取[4?5]。使用SVM函數(shù)進(jìn)行數(shù)據(jù)的提取過程如下:
式中:[Ka,b]為提取后的數(shù)據(jù)統(tǒng)一表達(dá)方法;[x×y]為圖像的畫面幀,每一幀的畫面都是通過一定的數(shù)據(jù)變化進(jìn)行提取出來,因此只需要對不同幀進(jìn)行限定便可以進(jìn)行有效的提取。經(jīng)過上述過程完成了數(shù)據(jù)的提取。
1.2 遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性表示
本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,經(jīng)過對圖像數(shù)據(jù)的有效提取[6?7],能夠進(jìn)行數(shù)據(jù)的加權(quán)表示,首先是款量數(shù)據(jù)的加權(quán),過程如下:
式中:[Msq]為圖像的加權(quán)辨別數(shù)據(jù)的辨別集合;[pqi]為提取的辨別參考可用數(shù)據(jù);[εsq]為加權(quán)系數(shù)。
2 遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判別
2.1 引入快速魯棒識(shí)別算法endprint
本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,已經(jīng)使用權(quán)值對遠(yuǎn)程圖像的數(shù)據(jù)進(jìn)行了表示,通過本文的參考引入快速魯棒識(shí)別算法對遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判別。但是使用快速魯棒識(shí)別算法需要把加權(quán)數(shù)據(jù)進(jìn)行一次數(shù)據(jù)的預(yù)處理[6,8],這樣才能保證數(shù)據(jù)的有效性以及計(jì)算的準(zhǔn)確度。預(yù)處理過程如下:
式中:[M2ij]為快速魯棒識(shí)別算法中的使用參比數(shù)據(jù);[Mi],[Mj]分別為款量數(shù)據(jù)的加權(quán)系數(shù)、識(shí)別數(shù)據(jù)的加權(quán)系數(shù);[Qij]為數(shù)據(jù)的最終變量差;[sinεij]為多維數(shù)的影響對照參數(shù);[P2ij]為高階變量參數(shù)。使用統(tǒng)一的魯棒配比才能進(jìn)行效果表達(dá),過程如下:
式中:[πik]為數(shù)據(jù)有效魯棒性;[MK]為參比量使用數(shù)據(jù)的高階閾值;[QiK]為常規(guī)的數(shù)據(jù)跳躍度;[linMsε]為數(shù)據(jù)變量預(yù)備值。經(jīng)過上述過程完成了對數(shù)據(jù)的有效預(yù)處理。
2.2 完成遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷
本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,經(jīng)過上述過程的預(yù)處理后便可以進(jìn)行有效的辨別,快速魯棒識(shí)別算法首選對魯棒系數(shù)進(jìn)行表達(dá)確認(rèn),過程如下:
3 仿真實(shí)驗(yàn)分析
3.1 參數(shù)設(shè)定
為了保證本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法的有效性,對參數(shù)進(jìn)行設(shè)置,設(shè)置加權(quán)過后的款量數(shù)據(jù)在值域[15.5,18.5]以內(nèi),設(shè)置[Mi],[M2ij],[Sj],[Qij]分別為15,12.5,600,523。
本文設(shè)計(jì)的實(shí)驗(yàn)遠(yuǎn)程圖像數(shù)據(jù)的選定過程中特意地對不同數(shù)據(jù)進(jìn)行了有效的分離,這樣能夠更加直觀地對設(shè)計(jì)的有效性進(jìn)行觀察。數(shù)據(jù)分布如圖1所示。
3.2 節(jié)能數(shù)據(jù)誤差調(diào)節(jié)
為了保證本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法的有效性,需要對實(shí)驗(yàn)數(shù)據(jù)進(jìn)行參設(shè),其參設(shè)的數(shù)據(jù)如表1所示。
3.3 結(jié)果對比分析
分析圖2結(jié)果得知,本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,在判別效率上明顯的要高于傳統(tǒng)的判別方法,同時(shí)判別時(shí)間是傳統(tǒng)方法的[12]左右。
分析圖3結(jié)果得知,通過與設(shè)定的標(biāo)準(zhǔn)精度進(jìn)行對比,能夠看出本文設(shè)計(jì)的基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法幾乎能夠與理論值相吻合,沒有較大的判斷誤差。
4 結(jié) 語
本文提出一種基于遠(yuǎn)程圖像識(shí)別的景觀設(shè)計(jì)合理性判斷方法,并通過權(quán)值對比有效地對遠(yuǎn)程圖像識(shí)別景觀設(shè)計(jì)合理性進(jìn)行判斷。實(shí)驗(yàn)結(jié)果表明,采用改進(jìn)的判斷方法,準(zhǔn)確性較高。
參考文獻(xiàn)
[1] 吳建勝,祖旭良,薛涵今.基于特征字符識(shí)別的文本圖像方向判定[J].計(jì)算機(jī)工程與設(shè)計(jì),2015(8):2260?2263.
WU Jiansheng, ZU Xuliang, XUE Hanjin. Determining direction of text image based on feature character recognition [J]. Computer engineering and design, 2015(8): 2260?2263.
[2] 董薇.紫金山地區(qū)盜伐立木三維圖像特征的遠(yuǎn)程可視化研究[J].科技通報(bào),2016,32(10):148?151.
DONG Wei. Remote visualization of three?dimensional image of standing purple with illegal logging feature in Zijinshan Mountain area [J]. Bulletin of science and technology, 2016, 32(10): 148?151.
[3] 彭贊,蔣外文,羅安,等.基于圖像智能識(shí)別技術(shù)的遠(yuǎn)程視頻監(jiān)控系統(tǒng)[J].桂林電子科技大學(xué)學(xué)報(bào),2004,24(5):23?26.
PENG Zan, JIANG Waiwen, LUO An, et al. A long?range video surveillance system based on digital image recognition technique [J]. Journal of Guilin University of Electronic Technology, 2004, 24 (5): 23?26.
[4] 羅會(huì)銘,彭贊,李正國.一種基于智能圖像識(shí)別的遠(yuǎn)程視頻監(jiān)控系統(tǒng)[J].計(jì)算技術(shù)與自動(dòng)化,2005,24(2):51?53.
LUO Huiming, PENG Zan, LI Zhengguo. A remote video monitoring system based on intelligent image recognition [J]. Computing technology and automation, 2005, 24(2): 51?53.
[5] 李雄軍,郭應(yīng)龍,肖煥雄.鐵路景觀綠化設(shè)計(jì)探析[J].微型機(jī)與應(yīng)用,1997,6(8):49?50.
LI Xiongjun, GUO Yinglong, XIAO Huanxiong. A study on the landscape design of railway line [J]. Microcomputers and applications, 1997, 6(8): 49?50.
[6] 郭秦龍,聞立杰,金濤,等.NBAJ:一種基于網(wǎng)絡(luò)流的工作流資源分配合理性判定方法[J].計(jì)算機(jī)集成制造系統(tǒng),2015,21(2):326?335.
GUO Qinlong, WEN Lijie, JIN Tao, et al. NBAJ: workflow resource assignment satisfaction judgment method based on network?flow [J]. Computer integrated manufacturing systems, 2015, 21(2): 326?335.
[7] 劉英憲,馬奎前,穆朋飛,等.一種水驅(qū)曲線直線段合理性判斷新方法[J].特種油氣藏,2016,23(5):86?88.
LIU Yingxian, MA Kuiqian, MU Pengfei, et al. A new method of rational judgment of straight line section of water drive curve [J]. Special oil and gas reservoirs, 2016, 23(5): 86?88.
[8] 秦國輝.基于三維圖像的運(yùn)動(dòng)合理性判斷方法研究與仿真[J]. 現(xiàn)代電子技術(shù),2017,40(8):22?24.
QIN Guohui. Study and simulation of motion rationality judgment method based on 3D image [J]. Modern electronics technique, 2017, 40(8): 22?24.
[9] 朱虹,楊登輝.一種基于圖像識(shí)別與RFID技術(shù)的貨車進(jìn)出管理系統(tǒng)[J].物聯(lián)網(wǎng)技術(shù),2016,6(7):68?69.
ZHU Hong, YANG Denghui. A truck entry and exit mangement system based on image recognition and RFID technology [J]. Internet of Things technologies, 2016, 6(7): 68?69.
[10] 王偉,王斌,孫俊蘭.基于兩層魯棒性交叉映射的大數(shù)據(jù)融合算法[J].計(jì)算機(jī)仿真,2016,33(10):236?239.
WANG Wei, WANG Bin, SUN Junlan. Based on two layers of cross mapping large data fusion algorithm robustness [J]. Computer simulation, 2016, 33(10): 236?239.endprint