探地雷達(dá)多目標(biāo)識(shí)別方法的研究
Research on GPR Multi-object Recognition
-
摘要: 與現(xiàn)有的機(jī)器學(xué)習(xí)算法相比，在有限樣本的情況下，支撐矢量機(jī)具有更強(qiáng)的分類(lèi)推廣能力。該文在提出利用非線性映射進(jìn)行探地雷達(dá)目標(biāo)識(shí)別的基礎(chǔ)上，將多目標(biāo)識(shí)別支撐矢量機(jī)與探地雷達(dá)目標(biāo)識(shí)別相結(jié)合，得到了基于一對(duì)一(One against one) 支撐矢量機(jī)的探地雷達(dá)多目標(biāo)識(shí)別方法。所提方法包括基于一對(duì)一的探地雷達(dá)多目標(biāo)識(shí)別方法、交叉驗(yàn)證的參數(shù)選取方法、多通道識(shí)別方法；并且和傳統(tǒng)的神經(jīng)網(wǎng)絡(luò)識(shí)別方法進(jìn)行對(duì)比分析。所提識(shí)別方法可以與各種目標(biāo)特征選取方法相結(jié)合。對(duì)實(shí)測(cè)數(shù)據(jù)的對(duì)比處理表明所提方法優(yōu)于傳統(tǒng)探地雷達(dá)目標(biāo)識(shí)別方法，所得結(jié)論對(duì)探地雷達(dá)目標(biāo)識(shí)別的研究有指導(dǎo)意義。Abstract: With limited samples, SVM has stronger ability of generalization in comparison with machine learning algorithm. In this paper, the SVM is combined with the Ground Penetrating Radar(GPR) multi-object recognition, and a GPR multi-object recognition method is proposed based on the one against one SVM. The proposed method includes the GPR multi-object recognition method based on one against one SVM, the parameter selection method based on the cross-validation and the multichannel recognition method. The contrast analysis between the proposed method and the conventional neural network method is given. The proposed method can be combined with object-feature extraction methods. It is shown that the method is effective in the experimental analysis. The conclusion can direct the research on GPR object recognition.
-
計(jì)量
- 文章訪問(wèn)數(shù): 2293
- HTML全文瀏覽量: 78
- PDF下載量: 1189
- 被引次數(shù): 0
下載:
