1-Bit壓縮感知盲重構(gòu)算法
doi: 10.11999/JEIT140419 cstr: 32379.14.JEIT140419
基金項(xiàng)目:
國家自然科學(xué)基金(61102148)和黑龍江省博士后基金(LBH-Z10167)資助課題
A Blind 1-Bit Compressive Sensing Reconstruction Method
-
摘要: 1-Bit壓縮感知(CS)是壓縮感知理論的一個(gè)重要分支。該領(lǐng)域中二進(jìn)制迭代硬閾值(BIHT)算法重構(gòu)精度高且一致性好，是一種有效的重構(gòu)算法。該文針對(duì)BIHT算法重構(gòu)過程需要信號(hào)稀疏度為先驗(yàn)信息的問題，提出一種稀疏度自適應(yīng)二進(jìn)制迭代硬閾值算法，簡(jiǎn)稱為SABIHT算法。該算法修正了BIHT算法，首先通過自適應(yīng)過程自動(dòng)調(diào)節(jié)硬閾值參數(shù)，然后利用測(cè)試條件估計(jì)信號(hào)的稀疏度，最終實(shí)現(xiàn)不需要確切信號(hào)稀疏度的1-Bit壓縮感知盲重構(gòu)。理論分析和仿真結(jié)果表明，該算法較好地實(shí)現(xiàn)了未知信號(hào)稀疏度的精確重建，并且與BIHT算法相比重構(gòu)精度及算法復(fù)雜度均相當(dāng)。
-
關(guān)鍵詞:
- 壓縮感知 /
- 1-Bit壓縮感知 /
- 盲重構(gòu) /
- 二進(jìn)制迭代硬閾值
Abstract: 1-Bit Compressive Sensing (CS) is an important branch of standard CS. The existing 1-Bit CS algorithm-Binary Iterative Hard Thresholding (BIHT) can perfectly recovery signals with high precision and consistency, which requires exact sparsity level in the recovery phase. Considering this problem, a new Sparsity Adaptive Binary Iterative Hard Thresholding (SABIHT) algorithm without prior information of the sparsity is proposed by modifying the BIHT algorithm. By using the adaptive process of automatically adjusting the hard threshold parameters and test conditions to estimate the sparsity, the proposed algorithm realizes accurate reconstruction and estimates the true supporting set of approximated signal. The analytical theory and simulation results show that the SABIHT algorithm recovers effectively the signals without prior information of signal sparsity and the reconstruction performance is similar to the BIHT algorithm. -
計(jì)量
- 文章訪問數(shù): 3005
- HTML全文瀏覽量: 253
- PDF下載量: 1006
- 被引次數(shù): 0
下載:
