基于混合矩的極化SAR圖像K分布模型參數(shù)估計(jì)新方法
doi: 10.11999/JEIT140551
基金項(xiàng)目:
國(guó)家自然科學(xué)基金(61372165)和湖北省自然科學(xué)基金(2012FB06902)資助課題
Parameter Estimation for the K-distribution in PolSAR Imagery Based on Hybrid Moments
-
摘要: K分布模型在極化合成孔徑雷達(dá)(PolSAR)圖像建模領(lǐng)域中獲得廣泛應(yīng)用。其模型參數(shù)估計(jì)的精度將直接影響到模型擬合的準(zhǔn)確性。目前普遍采用的K分布參數(shù)估計(jì)方法是基于協(xié)方差矩陣Mellin變換的對(duì)數(shù)累積量的估計(jì)方法。但是該方法沒有解析的表達(dá)式，數(shù)值計(jì)算運(yùn)算時(shí)間較長(zhǎng)，另外在形狀參數(shù)1時(shí)估計(jì)偏差較大。為此該文提出一種基于|z|rlg|z|混合矩的參數(shù)估計(jì)新方法，該方法對(duì)不同形狀參數(shù)值下的參數(shù)估計(jì)具有較好的適應(yīng)性，并且在值較小時(shí)估計(jì)性能優(yōu)于對(duì)數(shù)累積量方法。同時(shí)在r=1/d時(shí)該方法有解析的表達(dá)式，其運(yùn)算時(shí)間優(yōu)于對(duì)數(shù)累積量方法。最后用仿真數(shù)據(jù)和實(shí)測(cè)數(shù)據(jù)對(duì)新方法和已有參數(shù)估計(jì)方法的結(jié)果進(jìn)行了比較，驗(yàn)證了基于混合矩估計(jì)方法的準(zhǔn)確性與有效性。該方法為PolSAR圖像統(tǒng)計(jì)模型參數(shù)的快速有效估計(jì)提供了新手段。
-
關(guān)鍵詞:
- 極化合成孔徑雷達(dá) /
- 混合矩 /
- K分布 /
- 參數(shù)估計(jì)
Abstract: The K-distribution is usually used to model the Polarimetric Synthetic Aperture Radar (PolSAR) imagery. The parameter estimation method for K-distribution is very important,which affects the fitting degree of the model. However, the classical method of matrix log-cumulants relies upon a nontrivial inversion of a nonlinear equation, which introduces a computationally expensive stage into the estimation procedure. Moreover, the bias is large when the sharp parameter is smaller than 1. Therefore, a new method for estimating the sharp parameter of K-distribution based on|z|rlg|z| is proposed. This method is more adaptable to parameter estimation under different sharp parameters, and the performance is better than matrix log-cumulantes whenis small. In addition, the proposed estimator has an analytical expression at r=1/d, which allows rapid caculation. Finally, the estimation accuracy of this new estimator is compared with previous ones through simulation data and real data. The results show that the new estimator is effective and robust, which is of practical value in solving the accurate parameter estimation of K-distribution. -
計(jì)量
- 文章訪問數(shù): 1676
- HTML全文瀏覽量: 131
- PDF下載量: 554
- 被引次數(shù): 0
下載:
