百度语音识别极速版的使用例子（JAVA）

说明：
1，百度语音极速版，需要事先保存声音文件，然后调用百度接口。

2，关于文件的上传，可以用JSON格式，上传文件的base64编码，也可以用row格式直接上传文件。下面的代码是JSON格式。

3，使用接口前需要向百度AI开放平台注册APPID，得到秘钥，然后用得到的APPID和秘钥来获得TOKEN。调用百度语音接口时需要TOKEN。

4，百度提供了一个工具，可以从拼音角度判断相似度，score越小说明拼音越相似，挺好用，非常适合需要匹配的场景。

5，相关Maven库：

<dependency><groupId>com.github.stuxuhai</groupId><artifactId>jpinyin</artifactId><version>1.1.8</version>
</dependency>

上代码：

package test;import com.alibaba.fastjson.JSONObject;
import sun.misc.BASE64Encoder;import java.io.*;
import java.net.HttpURLConnection;
import java.net.URL;
import java.util.List;/*** 百度语音识别相关工具*/
public class VopUtil {public final static String URL = "https://vop.baidu.com/pro_api";public final static String TOKEN = "";  //需要事先注册public final static String CUID = "zheshiyigeceshi"; //用户唯一标识，用来区分用户，计算UV值。建议填写能区分用户的机器 MAC 地址或 IMEI 码，长度为60字符以内。public static void main(String[] args){String filePath="D:\\语音test\\我要买美汁源果粒橙.wav";    //语音文件路径，这个文件是我自己用手机的录音机录的一段语音，内容就是文件名String base64=getBaset64(filePath);                       //文件转为base64格式long fileLength=getFileLength(filePath);                  //获得文件长度，单位字节JSONObject vopResult = getVopStr(base64, fileLength);     //调用百度语音识别接口System.out.println("语音识别接口返回值："+vopResult.toJSONString());String errMsg = vopResult.get("err_msg").toString();if (!"success.".equals(errMsg)) {System.out.println("识别失败");}String resultStr = vopResult.get("result").toString();resultStr = resultStr.replaceAll("\\[", "").replaceAll("]", "").replaceAll("\"", "");System.out.println("语音识别的结果："+resultStr);String[] nameArr={"可乐","雪碧","美年达","美汁源","红牛","尖叫","牛奶","酸奶","百事"};  //备选名字列表，用于匹配拼音相似度String targetName = getSimilarWord(resultStr, nameArr);System.out.println("拼音最相似："+targetName);}/*** 请求百度语音识别接口* @param file 文件的base64编码* @param len 文件长度，字节数* @return*/public static JSONObject getVopStr(String file,long len) {JSONObject params = new JSONObject(true);params.put("format", "wav");params.put("rate", 16000);params.put("channel", 1);params.put("cuid", CUID);params.put("token", TOKEN);params.put("dev_pid", 80001);params.put("len", len);params.put("speech", file);String result = "";BufferedReader reader = null;BufferedReader in = null;try {URL url = new URL(URL);// 创建连接HttpURLConnection connection = (HttpURLConnection) url.openConnection();connection.setDoOutput(true);connection.setDoInput(true);connection.setUseCaches(false);connection.setInstanceFollowRedirects(true);connection.setRequestMethod("POST"); // 设置请求方式connection.setRequestProperty("Accept", "application/json"); // 设置接收数据的格式connection.setRequestProperty("Content-Type", "application/json"); // 设置发送数据的格式connection.connect();if (params != null) {// 设置文件长度//   connection.setRequestProperty("Content-Length", String.valueOf(writebytes.length));OutputStream outwritestream = connection.getOutputStream();outwritestream.write(params.toJSONString().getBytes());outwritestream.flush();outwritestream.close();}int responseCode = connection.getResponseCode();InputStream inputStream = null;if (responseCode == 200) {inputStream = new BufferedInputStream(connection.getInputStream());} else {inputStream = new BufferedInputStream(connection.getErrorStream());}in = new BufferedReader(new InputStreamReader(inputStream, "UTF-8"));String line;while ((line = in.readLine()) != null) {result += line;}} catch (Exception e) {e.printStackTrace();} finally {if (reader != null) {try {reader.close();} catch (IOException e) {e.printStackTrace();}}}return JSONObject.parseObject(result);}/*** 从目标字符串列表中查询拼音最相似的字符串* @param input 待查询的字符串* @param arr 备选词语库* @return*/public static String getSimilarWord(String input,String[] arr){try {Search d = new Search(arr);List<Search.Score> list=d.search(input,10);System.out.println("拼音相似度列表："+list);if(list!=null&&list.size()>0){Search.Score score=list.get(0);return score.word.toString();}}catch (Exception e){e.printStackTrace();}return null;}/*** 文件转base64* @param filePath 文件路径* @return*/private static String getBaset64(String filePath){String base64="";try {File file = new File(filePath);FileInputStream inputFile = new FileInputStream(file);byte[] buffer = new byte[(int)file.length()];inputFile.read(buffer);inputFile.close();base64= new BASE64Encoder().encode(buffer);if(base64.contains("\n")){base64=base64.replaceAll("\n","");}if(base64.contains("\r")){base64=base64.replaceAll("\r","");}}catch (Exception e){e.printStackTrace();}return base64;}/*** 得到文件大小，单位：字节* @param filePath 文件路径* @return*/private static long getFileLength(String filePath){File file = new File(filePath);long fileLength=file.length();return fileLength;}}

下面是Search类，这个类是百度提供的：

package test;import com.github.stuxuhai.jpinyin.PinyinException;
import com.github.stuxuhai.jpinyin.PinyinFormat;
import com.github.stuxuhai.jpinyin.PinyinHelper;import java.util.ArrayList;
import java.util.List;
import java.util.stream.Collectors;/*** 本文件用来演示如何通过文本相似度进行纠错* <p>* <pre>* String[] list = new String[]{"张三", "张衫", "张丹", "张成", "李四", "李奎"};* Search d = new Search(list);* System.out.println(d.search("张三", 10));* System.out.println(d.search("李四", 10));** 输出：* [{word=张三, score=0}, {word=张衫, score=1}, {word=张丹, score=1}, {word=张成, score=5}, {word=李四, score=9}, {word=李奎, score=10}]* [{word=李四, score=0}, {word=李奎, score=3}, {word=张三, score=9}, {word=张衫, score=10}, {word=张丹, score=10}, {word=张成, score=12}]* </pre>*/
public class Search {final List<Word> targets = new ArrayList<Word>();public Search(String[] list) throws PinyinException {for (String s : list) {Word w = new Word(s);targets.add(w);}}public List<Score> search(String input, int limit) throws PinyinException {Word w = new Word(input);return targets.stream().map(x -> {Score s = new Score();s.word = x;s.score = x.compareTo(w);return s;}).sorted().limit(limit).collect(Collectors.toList());}public static int getEditDistance(String s, String t) {int d[][]; // matrixint n; // length of sint m; // length of tint i; // iterates through sint j; // iterates through tchar s_i; // ith character of schar t_j; // jth character of tint cost; // cost// Step 1n = s.length();m = t.length();if (n == 0) {return m;}if (m == 0) {return n;}d = new int[n + 1][m + 1];// Step 2for (i = 0; i <= n; i++) {d[i][0] = i;}for (j = 0; j <= m; j++) {d[0][j] = j;}// Step 3for (i = 1; i <= n; i++) {s_i = s.charAt(i - 1);// Step 4for (j = 1; j <= m; j++) {t_j = t.charAt(j - 1);// Step 5cost = (s_i == t_j) ? 0 : 1;// Step 6d[i][j] = Minimum(d[i - 1][j] + 1, d[i][j - 1] + 1,d[i - 1][j - 1] + cost);}}// Step 7return d[n][m];}private static int Minimum(int a, int b, int c) {int im = a < b ? a : b;return im < c ? im : c;}class Word implements Comparable {final String word;final String pinyin1;final String pinyin2;Word(String word) throws PinyinException {this.word = word;this.pinyin1 = PinyinHelper.convertToPinyinString(word, ",", PinyinFormat.WITH_TONE_NUMBER);this.pinyin2 = PinyinHelper.convertToPinyinString(word, ",", PinyinFormat.WITHOUT_TONE);}@Overridepublic String toString() {return word;}@Overridepublic int compareTo(Object o) {if (o instanceof Word) {Word o1 = (Word) o;int score1 = getEditDistance(this.pinyin1, o1.pinyin1);int score2 = getEditDistance(this.pinyin2, o1.pinyin2);return score1 + score2;}return 0;}}class Score implements Comparable {Word word;int score;@Overridepublic int compareTo(Object o) {if (o instanceof Score) {return score - ((Score) o).score;}return 0;}@Overridepublic String toString() {return "{" +"word=" + word +", score=" + score +'}';}}
}

本地执行main方法后的输出结果：

语音识别接口返回值：{"result":["我要买美汁源果粒橙。"],"err_msg":"success.","sn":"275665891801562141461","corpus_no":"6709346487565939628","err_no":0}
语音识别的结果：我要买美汁源果粒橙。
拼音相似度列表：[{word=美汁源, score=58}, {word=美年达, score=66}, {word=百事, score=73}, {word=红牛, score=74}, {word=尖叫, score=74}, {word=酸奶, score=75}, {word=可乐, score=76}, {word=牛奶, score=76}, {word=雪碧, score=77}]
拼音最相似：美汁源

百度语音识别极速版的使用例子（JAVA）相关推荐

react-native 使用百度语音识别极速版语音转文字
项目有需求,需要集成语音转文字,于是开始研究,一步一个坑 1.集成录音 react-native-audio 2.集成文件操作 react-native-fs 3.集成音频格式转换 react-nat ...
天猫精灵 python 控制_树莓派4B小爱同学、天猫精灵、智能音箱、百度语音极速版、百度语音标准版、语音识别、语音合成终极方案...
来源:树莓派创客空间 https:/r-pi.cn,里面更多教程第一句话,python2版本的代码,python3你改改就完事了第一句话,python2版本的代码,python3你改改就完事了流 ...
树莓派小爱同学、天猫精灵、智能音箱、百度语音极速版、百度语音标准版、语音识别、语音合成终极方案
第一句话,python2版本的代码,python3你改改就完事了流程图如下: 全套教程下载地址点击下载首先需要调用这个方法安装百度的语音识别 https://ai.baidu.com/ai-doc ...
百度大脑语音识别极速版~C#开荒攻略
作者:goJhou 语音识别极速版 API 将60秒以内的完整音频文件识别为文字,适用于近场短语音交互,如手机语音搜索.聊天输入等场景.输入完整音频文件,输出识别结果文字. 采用流式多级截断注意力模型 ...
百度地图－极速版搭建手册
1.百度地图极速版简介本套API是专为手机浏览器提供的API,您可使用该套API,在手机页面中展示地图.标注位置.检索poi.查询线路等. 特点: 较同时兼容PC和手机浏览器的JavaScript ...
python调用百度语音搜索_使用 Python 和百度语音识别生成视频字幕
从视频中提取音频安装 moviepy pip install moviepy 相关代码: audio_file = work_path + '\\out.wav' video = VideoFile ...
使用 Python 和百度语音识别生成视频字幕
文章目录从视频中提取音频根据静音对音频分段使用百度语音识别获取 Access Token 使用 Raw 数据进行合成生成字幕总结从视频中提取音频安装 moviepy pip insta ...
python怎么加字幕_使用Python和百度语音识别生成视频字幕的实现
从视频中提取音频安装 moviepy pip install moviepy 相关代码: audio_file = work_path + '\\out.wav' video = VideoFile ...
python 视频语音转文字_使用Python和百度语音识别生成视频字幕的实现
从视频中提取音频安装 moviepy pip install moviepy 相关代码: audio_file = work_path + '\\out.wav' video = VideoFile ...

百度语音识别极速版的使用例子（JAVA）

百度语音识别极速版的使用例子（JAVA）相关推荐

最新文章

热门文章