Abstract: This paper does much labeling according to the law of the Uyghur’s accent location, and extracts different characteristic parameters（energy, pitch and so on）, does single flow, triage and feature fusion recognition to each characteristic parameters and compares analysis of the characteristic parameters to accent detection and identification rate. The feature after the integration of high dimensional flow characteristics using PCA is done dimensionality reduction and the redundant processing, and the recognition experiments are done. In the light of the recognition accuracy rate, combined with speech linguistic knowledge the experimental results are analysed.

Key words: Uyghur, accent detection, feature fusion, Principal Component Analysis（PCA）

摘要： 根据维吾尔语词重音的位置规律进行音节级标注，提取不同的特征参数（能量、基频等），对各个特征参数作单流、分流及特征级融合识别实验，对比分析各特征参数对重音检测识别率的影响。对融合后的高维单流特征采用主成分分析作降维、去冗余处理，并作识别实验。参照识别精确率结合语音语言学知识对实验结果进行分析。

关键词: 维吾尔语, 重音检测, 特征融合, 主成分分析