CAJ | 학술논문

频谱对比是曲线对比的一种,但由于频谱分布幅值差别大的特性,用传统的图像对比或匹配算法,运算量大,且精度不高。文章在分析阈值法以及最大互相关算法优缺点的基础上,提出了基于关键点的字符串对比算法,该算法采用按关键频率点进行分段,然后符号化曲线继而进行关键频率点的字符对比,算法能适应电磁扫描过程中的频谱对比要求,提高对比的精确性。
빈보대비시곡선대비적일충,단유우빈보분포폭치차별대적특성,용전통적도상대비혹필배산법,운산량대,차정도불고。문장재분석역치법이급최대호상관산법우결점적기출상,제출료기우관건점적자부천대비산법,해산법채용안관건빈솔점진행분단,연후부호화곡선계이진행관건빈솔점적자부대비,산법능괄응전자소묘과정중적빈보대비요구,제고대비적정학성。
Spectrum comparison is one type of curve comparison. Due to the great difference of spectral distribution amplitudes, traditional image comparison or matching algorithms have disadvantages such as large operation amount and low accuracy. Based on an analysis of the advantages and disadvantages of threshold method and maximum cross correlation algorithm, a string comparison algorithm based on key frequency points is proposed, which uses key frequency points for fragmentation, then symbolizes the curve, and compares the characters of key frequency points. The comparison algorithm is adaptable to the spectrum comparison requirement during electromagnetic scan, and improves the accuracy of comparison.