常方敏 李素敏



摘? 要: 針對最小二乘算法在處理坐標轉換存在準則單一這一問題,本文提出一種思想,將和聲搜索算法加入最小二乘坐標轉換模型中,從而使傳統坐標轉換這一P類問題,轉換為NP類問題。通過不斷迭代使其結果趨近于所設定的函數解,從而實現解的多樣性。本文以方向精度最優為例,分別采用模擬數據和實測點云數據來驗證本文思想。結果表明,在求某一方向精度時將和聲搜索算法加入模型中會提高方向精度,加權后也可以保證其垂直方向一定的精度。
關鍵詞: 和聲搜索算法;坐標轉換;方向精度;坐標轉換準則
中圖分類號: P228? ? 文獻標識碼: A? ? DOI:10.3969/j.issn.1003-6970.2020.10.027
本文著錄格式:常方敏,李素敏. 和聲搜索算法在坐標轉換中的應用[J]. 軟件,2020,41(10):109113
【Abstract】: Aiming at the problem that the least squares algorithm has a single criterion when dealing with coordinate transformation, this paper proposes an idea to add an intelligent optimization algorithm to the least square coordinate transformation model, so that the traditional polynomial problem of coordinate transformation is converted to the Non-polynomial problem. Through continuous iteration, the results approach the set function solution, so as to realize the diversity of the solution. In this paper, taking the best direction accuracy as an example, the simulation data and the measured point cloud data are used to verify the idea of this paper. The results show that HS algorithm is added to the model, the direction accuracy can be improved, and the vertical direction accuracy can be guaranteed after weighting.
【Key words】: Harmony search algorithm; Coordinate transformation; Direction accuracy; Coordinate transformation criterion
0? 引言
在2001年,Zong W G等[1]提出了一種新穎的啟發式智能優化算法——和聲搜索(Harmony Search,HS)算法,由于該算法操作簡單,收斂性與初始值無關,具備一定魯棒性并且可以避免陷入局部最優等優點,因此近年來HS應用在各個領域,例如:穩健大地測量[2]、地下水位的長期預測[3]、斜坡堤護面塊體[4]等等。
傳統的最小二乘(Least Square,LS)算法在用于坐標轉換模型存在一定的缺陷[5]。文獻[6]在坐標轉換模型中引入基準旋轉中心,解決轉換系數矩陣病態的問題。文獻[7]提出基于單位四元數的任意旋轉角度的三維坐標轉換,解決了大角度轉換的問題。文獻[8]采用選權迭代的總體最小二乘算法,通過降低粗差的權重,從而得到更高精度轉換參數。文獻[9]在加權總體最小二乘的坐標轉換模型中加入驗后估計,解決了加權總體最小二乘單位權方差不同的問題。……