Android百度语音识别/语音助手

发布时间：2024-11-20 17:39

AI语音识别技术应用于语音助手，辅助学生学习 #生活知识# #科技生活# #科技改变生活# #科技教育#

上一遍写了语音唤醒和语音合成，这篇写下语音合成

/** * 语音识别 * Created by fujiayi on 2017/6/13. * EventManager内的方法如send 都可以在主线程中进行，SDK中做过处理 */ public class RecognizerUtils { private final String TAG = this.getClass().getSimpleName(); private static RecognizerUtils utils = null; /** * SDK 内部核心 EventManager 类 */ private EventManager asr; // SDK 内部核心事件回调类，用于开发者写自己的识别回调逻辑 private EventListener eventListener; // 是否加载离线资源 private boolean isOfflineEngineLoaded = false; // 未release前，只能new一个 private volatile boolean isInited = false; /** * 0: 方案1， backTrackInMs > 0,唤醒词说完后，直接接句子，中间没有停顿。 * 开启回溯，连同唤醒词一起整句识别。推荐4个字 1500ms * backTrackInMs 最大 15000，即15s * * >0 : 方案2：backTrackInMs = 0，唤醒词说完后，中间有停顿。 * 不开启回溯。唤醒词识别回调后，正常开启识别。 * <p> * */ private static int backTrackInMs = 0; private static Handler handler = null; public static final int RECOGNIZER_ERROR = 1012; public static RecognizerUtils getIntance(){ if (utils == null){ synchronized (RecognizerUtils.class){ if (utils == null){ utils = new RecognizerUtils(); } } } return utils; } public void init(Handler mHandler){ handler = mHandler; if (isInited) { LogUtils.log(TAG,"还未调用release()，请勿新建一个新类"); return; } isInited = true; // SDK集成步骤初始化asr的EventManager示例，多次得到的类，只能选一个使用 asr = EventManagerFactory.create(MyApplication.context, "asr"); // SDK集成步骤设置回调event，识别引擎会回调这个类告知重要状态和识别结果 asr.registerListener(new EventListener() { @Override public void onEvent(String name, String params, byte[] data, int offset, int length) { eventListener = this; if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_LOADED)) { LogUtils.log("RecognizerUtils", "离线命令词资源加载成功"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_UNLOADED)) { LogUtils.log("RecognizerUtils", "离线命令词资源释放成功"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_READY)) { // 引擎准备就绪，可以开始说话 LogUtils.log("RecognizerUtils", "引擎准备就绪，可以开始说话"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_BEGIN)) { // 检测到用户的已经开始说话 LogUtils.log("RecognizerUtils", "检测到用户的已经开始说话"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_END)) { // 检测到用户的已经停止说话 LogUtils.log("RecognizerUtils", "检测到用户的已经停止说话"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_PARTIAL)) { RecogResult recogResult = RecogResult.parseJson(params); // 临时识别结果, 长语音模式需要从此消息中取出结果 LogUtils.log("RecognizerUtils", "临时识别结果, 长语音模式需要从此消息中取出结果"); String[] results = recogResult.getResultsRecognition(); if (recogResult.isFinalResult()) { LogUtils.log("RecognizerUtils", "isFinalResult 可能返回多个结果，请取第一个结果" + Arrays.toString(results)); Resolution.getIntance().resolution(handler, Arrays.toString(results)); } else if (recogResult.isPartialResult()) { LogUtils.log("RecognizerUtils", "onAsrBegin 后随着用户的说话，返回的临时结果" + Arrays.toString(results)); } else if (recogResult.isNluResult()) { LogUtils.log("RecognizerUtils", "isNluResult 可能返回多个结果，请取第一个结果" + new String(data, offset, length)); } } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_FINISH)) { // 识别结束，最终识别结果或可能的错误 LogUtils.log("RecognizerUtils", name + "- CALLBACK_EVENT_ASR_FINISH识别结束，最终识别结果或可能的错误"); RecogResult recogResult = RecogResult.parseJson(params); if (recogResult.hasError()) { // int errorCode = recogResult.getError(); // int subErrorCode = recogResult.getSubError(); if (handler != null){ handler.sendEmptyMessage(RECOGNIZER_ERROR); } LogUtils.log("RecognizerUtils", "识别结束，最终识别结果错误" + recogResult.getDesc()); } else { LogUtils.log("RecognizerUtils", "识别结束，最终识别结果" + params); } } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_LONG_SPEECH)) { // 长语音 LogUtils.log("RecognizerUtils", "长语音"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_EXIT)) { LogUtils.log("RecognizerUtils", "退出"); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_VOLUME)) { // Logger.info(TAG, "asr volume event:" + params); LogUtils.log("RecognizerUtils", params); } else if (name.equals(SpeechConstant.CALLBACK_EVENT_ASR_AUDIO)) { // LogUtils.log("RecognizerUtils", "长语音"); } } }); loadOfflineEngine(fetchOfflineParams()); } /** * 离线命令词，在线不需要调用 * * @param params 离线命令词加载参数，见文档“ASR_KWS_LOAD_ENGINE 输入事件参数” */ private void loadOfflineEngine(Map<String, Object> params) { String json = new JSONObject(params).toString(); // SDK集成步骤（可选）加载离线命令词(离线时使用) asr.send(SpeechConstant.ASR_KWS_LOAD_ENGINE, json, null, 0, 0); isOfflineEngineLoaded = true; // 没有ASR_KWS_LOAD_ENGINE这个回调表试失败，如缺少第一次联网时下载的正式授权文件。 } private Map<String, Object> fetchOfflineParams() { Map<String, Object> map = new HashMap<String, Object>(); map.put(SpeechConstant.DECODER, 2); map.put(SpeechConstant.ASR_OFFLINE_ENGINE_GRAMMER_FILE_PATH, "assets:///baidu_speech_grammar.bsg"); // map.putAll(fetchSlotDataParam()); return map; } private Map<String, Object> fetchSlotDataParam() { Map<String, Object> map = new HashMap<String, Object>(); try { JSONObject json = new JSONObject(); json.put("name", new JSONArray().put("妈妈").put("老伍")) .put("appname", new JSONArray().put("手百").put("度秘")); map.put(SpeechConstant.SLOT_DATA, json); } catch (JSONException e) { e.printStackTrace(); } return map; } /** * 开始 */ public void start() { // 此处开始正常识别流程 Map<String, Object> params = new LinkedHashMap<String, Object>(); if (isOfflineEngineLoaded){ params.put(SpeechConstant.DECODER, 2); } params.put(SpeechConstant.ACCEPT_AUDIO_VOLUME, false); params.put(SpeechConstant.VAD, SpeechConstant.VAD_DNN); // 如识别短句，不需要需要逗号，使用1536搜索模型。其它PID参数请看文档 params.put(SpeechConstant.PID, 1536); if (backTrackInMs > 0) { // 方案1 唤醒词说完后，直接接句子，中间没有停顿。开启回溯，连同唤醒词一起整句识别。 // System.currentTimeMillis() - backTrackInMs , 表示识别从backTrackInMs毫秒前开始 params.put(SpeechConstant.AUDIO_MILLS, System.currentTimeMillis() - backTrackInMs); } cancel(); // SDK集成步骤拼接识别参数 String json = new JSONObject(params).toString(); asr.send(SpeechConstant.ASR_START, json, null, 0, 0); } /** * 提前结束录音等待识别结果。 */ public void stop() { // SDK 集成步骤（可选）停止录音 if (!isInited) { // throw new RuntimeException("release() was called"); } if (asr == null) { return; } asr.send(SpeechConstant.ASR_STOP, "{}", null, 0, 0); } /** * 取消本次识别，取消后将立即停止不会返回识别结果。 * cancel 与stop的区别是 cancel在stop的基础上，完全停止整个识别流程， */ public void cancel() { if (!isInited) { // throw new RuntimeException("release() was called"); } if (asr == null) { return; } // SDK集成步骤 (可选）取消本次识别 asr.send(SpeechConstant.ASR_CANCEL, "{}", null, 0, 0); } public void release() { isInited = false; cancel(); if (asr == null) { return; } if (isOfflineEngineLoaded) { // SDK集成步骤如果之前有调用过加载离线命令词，这里要对应释放 asr.send(SpeechConstant.ASR_KWS_UNLOAD_ENGINE, null, null, 0, 0); isOfflineEngineLoaded = false; } // SDK 集成步骤（可选），卸载listener asr.unregisterListener(eventListener); asr = null; isInited = false; utils = null; } }

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223

/** * 解析语音识别 */ public class Resolution { private final String TAG = this.getClass().getSimpleName(); private static Resolution utils = null; public static Resolution getIntance() { if (utils == null) { synchronized (Resolution.class) { if (utils == null) { utils = new Resolution(); } } } return utils; } /** * 解析通过handler返回状态 * @param handler * @param work 识别结果 */ public void resolution(Handler handler, String work) { try { if (PhoneUtils.CALL_PHONE && (work.equals("[确定]") || work.equals("[是的]")) && !StringUtils.isEmpty(PhoneUtils.PHONE)){ PhoneUtils.getInstance().callPhone(PhoneUtils.PHONE); }else if (work.indexOf("打开") != -1) { InstallAppUtils.getIntanste().openApp(work); } else if (work.indexOf("打电话") != -1) { PhoneUtils.getInstance().call(work); }else { TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_no_hear), true); } } catch (Exception e) { e.printStackTrace(); } } }

12345678910111213141516171819202122232425262728293031323334353637383940414243

/** * 安装的APP实例 * Created by pc20170521 on 2017-10-28. */ public class InstallAppInfo { private boolean check = false; private String pakecge; private Drawable icon; private String name; private ComponentName component; public InstallAppInfo(){ } public InstallAppInfo(Drawable d, String s, ComponentName cn){ icon = d; name = s; component = cn; } public Drawable getIcon() { return icon; } public void setIcon(Drawable icon) { this.icon = icon; } public String getName() { return name; } public void setName(String name) { this.name = name; } public ComponentName getComponent() { return component; } public void setComponent(ComponentName component) { this.component = component; } public boolean isCheck() { return check; } public void setCheck(boolean check) { this.check = check; } public String getPakecge() { return component.getPackageName(); } public void setPakecge(String pakecge) { this.pakecge = pakecge; } }

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061

/** * 获取安装的APP * Created by pc20170521 on 2017-10-28. */ public class InstallAppUtils { private final String TAG = this.getClass().getSimpleName(); private ArrayList<InstallAppInfo> appList = null; private static InstallAppUtils utils = null; public static InstallAppUtils getIntanste(){ if (utils == null){ synchronized (InstallAppUtils.class){ if(utils == null){ utils = new InstallAppUtils(); } } } return utils; } /*** * 获取安装的APP * @return */ public ArrayList<InstallAppInfo> getInstallApp(){ appList = new ArrayList<>(); PackageManager pkgMgt = MyApplication.context.getPackageManager(); Intent it = new Intent(Intent.ACTION_MAIN); it.addCategory(Intent.CATEGORY_LAUNCHER); List<ResolveInfo> ra = pkgMgt.queryIntentActivities(it,0); for(int i=0;i<ra.size();i++){ ActivityInfo ai = ra.get(i).activityInfo; String label = ai.loadLabel(pkgMgt).toString(); ComponentName c = new ComponentName(ai.applicationInfo.packageName,ai.name); InstallAppInfo item = new InstallAppInfo(ai.loadIcon(pkgMgt),label,c); appList.add(item); } return appList; } public ArrayList<InstallAppInfo> getAppList() { if (appList == null){ appList = getInstallApp(); } return appList; } /*** * 语音打开某个APP */ public void openApp(String name){ name = name.substring(2); LogUtils.log(TAG, name); boolean isHasApp = false; for (int i = 0; i < getAppList().size(); i++) { if (name.contains(appList.get(i).getName())){ TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_open_wx) + appList.get(i).getName()); try { Intent intent; PackageManager packageManager = MyApplication.context.getPackageManager(); intent = packageManager.getLaunchIntentForPackage(appList.get(i).getPakecge()); MyApplication.context.startActivity(intent); isHasApp = true; break; } catch (Exception e) { TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_no_wx) + appList.get(i).getName()); e.printStackTrace(); } } } if (!isHasApp){ TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_no_hear)); } } }

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081

/** * 获取电话本/打电话工具类 */ public class PhoneUtils { private final String TAG = this.getClass().getSimpleName(); private static PhoneUtils utils = null; public static boolean CALL_PHONE = false;//语音拨电话请求 public static String PHONE = "";//电话号码 private List<PhoneInfo> list; public static PhoneUtils getInstance(){ if (utils == null){ synchronized (PhoneUtils.class){ if (utils == null){ utils = new PhoneUtils(); } } } return utils; } /** * 获取电话本 * @return */ public List<PhoneInfo> getPhoneList() { list = new ArrayList<>(); Cursor cursor = MyApplication.context.getContentResolver().query(ContactsContract.CommonDataKinds.Phone.CONTENT_URI, new String[] { "display_name", "sort_key", "contact_id", "data1" }, null, null, null); // moveToNext方法返回的是一个boolean类型的数据 while (cursor.moveToNext()) { //读取通讯录的姓名 String name = cursor.getString(cursor .getColumnIndex(ContactsContract.CommonDataKinds.Phone.DISPLAY_NAME)); //读取通讯录的号码 String number = cursor.getString(cursor .getColumnIndex(ContactsContract.CommonDataKinds.Phone.NUMBER)); int Id = cursor.getInt(cursor.getColumnIndex(ContactsContract.CommonDataKinds.Phone.CONTACT_ID)); String Sortkey = getSortkey(cursor.getString(1)); PhoneInfo phoneInfo = new PhoneInfo(name, number,Sortkey,Id); list.add(phoneInfo); // LogUtils.log("电话联系人", name); } cursor.close(); return list; } private static String getSortkey(String sortKeyString) { String key =sortKeyString.substring(0,1).toUpperCase(); if (key.matches("[A-Z]")){ return key; }else return "#"; } /** * 语音识别并拨打电话 */ public void call(String name){ try { name = name.substring(3); LogUtils.log(TAG, name); if (name.contains("10086")) { PHONE = "10086"; name = "10086"; callTips(name); } else { boolean isPhone = false; List<PhoneInfo> list = PhoneUtils.getInstance().getPhoneList(); for (int i = 0; i < list.size(); i++) { if (name.contains(list.get(i).getName()) || name.contains(list.get(i).getNumber())) { PHONE = list.get(i).getNumber(); name = list.get(i).getName(); isPhone = true; break; } } if (isPhone) { callTips(name); } else { TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_no_phone)); } } } catch (Exception e) { TTSUtils.getInstance().speak(MyApplication.context.getString(R.string.wakeup_no_phone)); e.printStackTrace(); } } /** * 拨打电话 */ private void callTips(String name) { CALL_PHONE = true; String tips = MyApplication.context.getString(R.string.wakeup_find_phone) + name + MyApplication.context.getString(R.string.wakeup_find_phone1); TTSUtils.getInstance().speak(tips, true); ProgressDialog dialog = new ProgressDialog(MyApplication.context); dialog.setTitle(tips); dialog.setOnDismissListener(new DialogInterface.OnDismissListener() { @Override public void onDismiss(DialogInterface dialog) { dialog.dismiss(); } }); } /** * 拨打电话 * @param phone */ public void callPhone(String phone) { Uri uri = Uri.parse("tel:" + phone); Intent intent = new Intent(Intent.ACTION_CALL, uri); if (ActivityCompat.checkSelfPermission(MyApplication.context, Manifest.permission.CALL_PHONE) != PackageManager.PERMISSION_GRANTED) { ToastUtils.show(MyApplication.context, MyApplication.context.getString(R.string.phone_permision)); PermissionsUtils.getInstance().showSystemPermissionsSettingDialog(MyApplication.activity); return; } MyApplication.activity.startActivityForResult(intent, 1); PHONE = ""; CALL_PHONE = false; } }

/** * Created by fujiayi on 2017/6/24. */ public class RecogResult { private static final int ERROR_NONE = 0; private String origalJson; private String[] resultsRecognition; private String origalResult; private String sn; // 日志id，请求有问题请提问带上sn private String desc; private String resultType; private int error = -1; private int subError = -1; public static RecogResult parseJson(String jsonStr) { RecogResult result = new RecogResult(); result.setOrigalJson(jsonStr); try { JSONObject json = new JSONObject(jsonStr); int error = json.optInt("error"); int subError = json.optInt("sub_error"); result.setError(error); result.setDesc(json.optString("desc")); result.setResultType(json.optString("result_type")); result.setSubError(subError); if (error == ERROR_NONE) { result.setOrigalResult(json.getString("origin_result")); JSONArray arr = json.optJSONArray("results_recognition"); if (arr != null) { int size = arr.length(); String[] recogs = new String[size]; for (int i = 0; i < size; i++) { recogs[i] = arr.getString(i); } result.setResultsRecognition(recogs); } } } catch (JSONException e) { e.printStackTrace(); } return result; } public boolean hasError() { return error != ERROR_NONE; } public boolean isFinalResult() { return "final_result".equals(resultType); } public boolean isPartialResult() { return "partial_result".equals(resultType); } public boolean isNluResult() { return "nlu_result".equals(resultType); } public String getOrigalJson() { return origalJson; } public void setOrigalJson(String origalJson) { this.origalJson = origalJson; } public String[] getResultsRecognition() { return resultsRecognition; } public void setResultsRecognition(String[] resultsRecognition) { this.resultsRecognition = resultsRecognition; } public String getSn() { return sn; } public void setSn(String sn) { this.sn = sn; } public int getError() { return error; } public void setError(int error) { this.error = error; } public String getDesc() { return desc; } public void setDesc(String desc) { this.desc = desc; } public String getOrigalResult() { return origalResult; } public void setOrigalResult(String origalResult) { this.origalResult = origalResult; } public String getResultType() { return resultType; } public void setResultType(String resultType) { this.resultType = resultType; } public int getSubError() { return subError; } public void setSubError(int subError) { this.subError = subError; } }

网址：Android百度语音识别/语音助手 https://www.yuejiaxmz.com/news/view/158425

⬅️上一篇：深度学习语音识别方法概述与分析

➡️下一篇：语音识别合成助手

Android百度语音识别/语音助手

相关内容

随便看看

最新动态分享

热点动态分享

专题

推荐动态分享