Crested Ibis Chirp The Findings Of A Study Of Characteristics And Behaviors

Along with the economic development in recent years, people rely on environmentpollution is more and more serious, all kinds of various environmental problems causepeople to seek all range of comfortable environment is more and more small, thissituation has influence to the wild animals, wild animals protection problem becomemore urgent. Crested ibis chirp research at the national level to protect wild animals andtheir behavior, the relationship between the to understand the status of the daily life ofthe crested ibis, mental activities and physiological mechanism are of great help,because of the crested ibis chirp signal contains abundant biological information. So thispaper studies the crested ibis chirp recognition behavior of the system, to protect thecrested ibis research provides a reference only.The crested ibis speech recognition technology is still less, therefore, the researchof this paper is to build on the foundation of modern speech signal processing, throughthe existing knowledge of speech signal processing, the existing voice recognitiontechnology is studied. The crested ibis chirp recognition system is the main researchcontent of the crested ibis chirp of preprocessing, feature extraction and patternmatching three big framework, in this paper, the first point to improve the endpointdetection in the identification system (afis) preprocess are part, due to the traditionalendpoint detection method based on energy and based on autocorrelation in lowsignal-to-noise ratio and a sudden background noise, the testing error is bigger, in thispaper based on the analysis of animal singing has a good effect on the spectra method,the method based on language spectra matrix to find the voiced parts of speech.Through the experimental results show that language spectra method can in the case ofa sudden noise, can still find the dullness of endpoint accurately. In this paper, thesecond improved code book point pattern matching in the system design, code design is the purpose of the book recorded four behavior of crested ibis chirp training into themost optimal code book, make the minimum distortion, common to the speechrecognition system of codebook design method is LBG algorithm and some intelligentalgorithm, but there are easy to fall into a state of the global optimal solution. Thisarticle is on the basis of the improved artificial colony algorithm puts forward a newcodebook design algorithm, and changed the formula leads the bee search, yes the entiresearch process can adaptively adjust the global search and local search. This effectivelyavoid easy to fall into local optimum. At last, through the method of experimentalcontrast three used for codebook design, prove that the improved algorithm not onlyimproves the convergence speed is improved the recognition rate.