The support vector classifier is adopted to identify fault in the two types , and the grid search method based on cross - validation is chosen to determine model parameters 該模型將變壓器故障分為放電性和過熱性兩大類,通過統(tǒng)計(jì)分析尋求特征量區(qū)分類間的故障類型,采用支持向量機(jī)識別類內(nèi)的故障類型,利用基于交叉驗(yàn)證的網(wǎng)格搜索法來確定支持向量機(jī)的參數(shù)。
In the end , two simple object functions are built based on the results of parameter analysis , with which to search the optimum ratio of spans by the grid searching method . the static behavior is also compared between the optimized bridge and the basic bridge 最后,以參數(shù)分析為基礎(chǔ),采用結(jié)構(gòu)優(yōu)化思想,建立兩個(gè)簡單的目標(biāo)函數(shù),運(yùn)用網(wǎng)格搜索法尋求最優(yōu)跨徑比值,并對優(yōu)化后的橋梁與基準(zhǔn)橋受力進(jìn)行對比分析。
More specifically , this dissertation includes following contents : firstly , based on the analysis of h . 264 , the effect on coding efficiency by the key modules of h . 264 is analyzed . secondly , for the purpose of improving the efficiency of h . 264 encoder , we optimized the h . 264 arithmetic . especially fast motion estimation is improved by using hucmhgs ( hybrid unsymmetrical - cross multi - hexagon - grid search ) for integer pel motion estimation and using cbfps ( center biased fractional pel search ) for fractional pel motion estimation 主要工作包括: ( 1 )研究h . 264標(biāo)準(zhǔn),分析編碼器各模塊及其采用的技術(shù)細(xì)節(jié)的計(jì)算復(fù)雜度,尋求提高編碼速度的途經(jīng); ( 2 )對h . 264的jm82軟件進(jìn)行算法總體優(yōu)化處理,尤其對于編碼速度的瓶頸? ?運(yùn)動(dòng)估計(jì)進(jìn)行改進(jìn),采用不對稱多六邊形網(wǎng)格混合搜索算法hucmhgs進(jìn)行整數(shù)像素運(yùn)動(dòng)估計(jì),采用帶金字塔搜索模式的中心偏置分?jǐn)?shù)像素搜索算法cbfps進(jìn)行分?jǐn)?shù)像素運(yùn)動(dòng)估計(jì); ( 3 )結(jié)合tms320c6416的結(jié)構(gòu)特點(diǎn),在研究并行優(yōu)化實(shí)現(xiàn)方法的基礎(chǔ)之上,對h . 264算法的各個(gè)具體模塊進(jìn)行匯編語言編程實(shí)現(xiàn)。
In this paper , when the global optimal theory and monte - carlo technique are applied , i transformed the maximum likelihood frequency estimation problem of sinusoidal waves from multidimensional grid search to a closed expression . this makes it easy to estimate the frequency through optimal calculation 本文即是從解決以上的問題入手,通過結(jié)合運(yùn)用全局優(yōu)化理論和monte - carlo技術(shù),使多個(gè)正弦波的最大似然頻率估計(jì)問題由多維格形搜索變成了一個(gè)閉合的表達(dá)式,從而可以通過優(yōu)化計(jì)算來進(jìn)行頻率估計(jì)。
Most existing identification algorithms used in robot soccer vision system need to scan the image another time after all the pixels " color have been classified , which is time consuming . in order to save time , a seed - fill algorithm based on grid search is introduced in this paper . the new algorithm is much faster 針對足球機(jī)器人系統(tǒng)中視覺系統(tǒng)現(xiàn)有的識別算法普遍需要對已經(jīng)進(jìn)行了顏色分類的像素點(diǎn)進(jìn)行二次掃描,占用大量時(shí)間的缺點(diǎn),提出了基于網(wǎng)格查找的新方法,并且這種網(wǎng)格查找與像素點(diǎn)的顏色分類同時(shí)進(jìn)行,并不占用處理時(shí)間。