Java实现本地语音识别与热点检测详解

<dependencies>
    <!-- Vosk Java SDK -->
    <dependency>
        <groupId>org.vosk</groupId>
        <artifactId>vosk-java</artifactId>
        <version>0.3.34</version>
    </dependency>

    <!-- Alibaba Cloud SDK for Java -->
    <dependency>
        <groupId>com.alibaba.nls</groupId>
        <artifactId>nls-sdk-java</artifactId>
        <version>3.0.5</version>
    </dependency>
</dependencies>

import java.util.HashSet;
import java.util.Set;

public class HotSpotWords {
    public static final Set<String> HOT_WORDS = new HashSet<>(Set.of(
        "你好", "退出", "播放音乐", "打开设置", "天气怎么样", "帮我查一下"
    ));
}

import org.vosk.Model;
import org.vosk.Recognizer;
import org.vosk.LibVosk;

import javax.sound.sampled.*;
import java.io.File;
import java.io.IOException;

public class LocalASRWithHotSpot {

    // 检测是否为热点语音
    private static boolean isHotSpot(String text) {
        return HotSpotWords.HOT_WORDS.stream().anyMatch(text::contains);
    }

    public static void main(String[] args) throws Exception {
        // 初始化 Vosk 库
        LibVosk.setLogLevel(0); // 设置日志级别

        // 加载本地模型（路径根据实际修改）
        Model model = new Model("models/cn-small");

        // 打开音频文件
        File audioFile = new File("samples/audio.wav");
        AudioInputStream ais = AudioSystem.getAudioInputStream(audioFile);

        // 转换音频格式为 PCM 16bit 单声道 16kHz
        AudioFormat targetFormat = new AudioFormat(16000, 16, 1, true, false);
        AudioInputStream convertedStream = AudioSystem.getAudioInputStream(targetFormat, ais);

        // 创建识别器
        Recognizer recognizer = new Recognizer(model, 16000);

        byte[] buffer = new byte[4096];
        int bytesRead;

        System.out.println("开始本地语音识别...");

        while ((bytesRead = convertedStream.read(buffer)) >= 0) {
            if (bytesRead > 0) {
                recognizer.acceptWaveForm(buffer, bytesRead);
            }
        }

        String result = recognizer.finalResult();
        System.out.println("最终识别结果: " + result);

        // 热点检测
        if (isHotSpot(result)) {
            System.out.println("发现热点语音内容: " + result);
        } else {
            System.out.println("非热点语音内容");
        }

        // 关闭资源
        recognizer.close();
        convertedStream.close();
        ais.close();
    }
}

import com.alibaba.nls.client.protocol.asr.AsrPidCallBack;
import com.alibaba.nls.client.protocol.asr.SpeechTranscriber;
import com.alibaba.nls.client.protocol.asr.SpeechTranscriberResponse;

import java.io.File;
import java.util.concurrent.CountDownLatch;

public class AliyunASRIntegration {

    public static void main(String[] args) throws Exception {
        String accessKeyId = "<your-access-key-id>";
        String accessKeySecret = "<your-access-key-secret>";
        String appKey = "<your-appkey>";

        String audioFilePath = "path/to/audio.wav"; // 支持 wav/pcm/amr

        CountDownLatch latch = new CountDownLatch(1);

        SpeechTranscriber asrClient = new SpeechTranscriber(accessKeyId, accessKeySecret, appKey, new AsrPidCallBack() {
            @Override
            public void onRecognitionResultChange(SpeechTranscriberResponse response) {
                String result = response.getTranscript();
                System.out.println("实时识别结果：" + result);
            }

            @Override
            public void onRecognitionCompleted(SpeechTranscriberResponse response) {
                String finalResult = response.getTranscript();
                System.out.println("最终识别结果：" + finalResult);
                latch.countDown();
            }

            @Override
            public void onTaskFailed(SpeechTranscriberResponse response) {
                System.err.println("识别失败: " + response.getErrorMsg());
                latch.countDown();
            }
        });

        // 设置语言模型、采样率等参数（根据音频格式调整）
        asrClient.setFormat("pcm"); // 可改为 wav/amr
        asrClient.setSampleRate(16000); // 根据音频采样率调整

        // 启动识别任务
        File audioFile = new File(audioFilePath);
        asrClient.start(audioFile);

        latch.await(); // 等待识别完成
        asrClient.stop(); // 停止客户端
    }
}