java调用腾讯文字识别接口：行驶证识别

本教程是识别行驶证的正副页面，要识别不同的图片根据文档调用不同的对象

配套前端是笔者发布的上一篇文章：uniapp调用手机的拍照和上传功能

• 引入SDK
• OCR文字识别API
• 对接OCR车辆行驶证接口
• 解析返回数据进行数据封装

引入SDK

<dependency>
  <groupId>com.tencentcloudapi</groupId>
  <artifactId>tencentcloud-sdk-java-ocr</artifactId>
  <version>3.1.1153</version>
</dependency>

OCR文字识别API

由于官方提供了SDK，所以我直接使用，也可以自己通过HTTP请求来进行文字识别

获取自己腾讯云密钥，建议使用子账号，赋予ocr权限

在自己的配置文件中进行配置

tencent:
  sercretKey: xxx
  sercretValue: xxx

属性绑定

@Component
@ConfigurationProperties(prefix = "tencent")
@Data
public class TencentOcr {
    private String sercretKey;
    private String sercretValue;
}

控制器编写

笔者需要index来判断正反面

@PassToken
@PostMapping("/identityDriverLicense")
@Operation(summary = "识别行驶证", description = "识别行驶证")
public Result<Void> identityDriverLicense(MultipartFile file,Integer index){
    userService.identityDriverLicense(file,index);
    return Result.success();
}

业务层编写

@Override
public void identityDriverLicense(MultipartFile file, Integer index) {
    try {
        byte[] bytes = file.getBytes();
        String base64Img = Base64.getEncoder().encodeToString(bytes);
        ocrClient.getVehicleLicenseInfo(base64Img,index);
    } catch (IOException e) {
        log.info(e.getMessage());
    }
}

工具类编写

• JsonUtil:自定义JSON处理工具类
• LicenseUtil:自定义response处理工具类
• DriverLicenseFront：封装响应数据的对象
• DriverLicenseBack：封装响应数据的对象
• DriverLicense：自定义枚举，杜绝魔法数字
• getVehicleRegisterCert：获取车辆登记证书的接口

@Component
@Slf4j
public class OcrClient {
    private final TencentOcr tencentOcr;
    @Autowired
    public OcrClient(TencentOcr tencentOcr) {
        this.tencentOcr = tencentOcr;
    }
    public Object getVehicleLicenseInfo(String base64Img,Integer index){
        try {
            VehicleLicenseOCRResponse response = performOcrRequest(base64Img, index);
            String json = JsonUtil.objectToJson(response);
            if (index == DriverLicense.FRONT.type()){
                DriverLicenseFront driverLicenseFront = LicenseUtil.getVehicleFront(json);
                return driverLicenseFront;
            } else if (index == DriverLicense.BACK.type()) {
                DriverLicenseBack vehicleInfoBack = LicenseUtil.getVehicleInfoBack(json);
                return vehicleInfoBack;
            }
        } catch (TencentCloudSDKException e) {
            log.info(e.getMessage());
            throw new GlobalException(504,"请选择正确的行驶证正副页");
        }
        return null;
    }
    public Object getVehicleRegisterCert(String base64Img){
        try {
            com.tencentcloudapi.ocr.v20181119.OcrClient client = createClient();
            VehicleRegCertOCRRequest request = new VehicleRegCertOCRRequest();
            request.setImageBase64(base64Img);
            VehicleRegCertOCRResponse response = client.VehicleRegCertOCR(request);
            String json = JsonUtil.objectToJson(response);
            System.out.println(json);
        } catch (TencentCloudSDKException e) {
            throw new GlobalException(504,"请选择有效的车辆登记合格证书");
        }
        return null;
    }
    private VehicleLicenseOCRResponse performOcrRequest(String base64Img, Integer index) throws TencentCloudSDKException {
        com.tencentcloudapi.ocr.v20181119.OcrClient client = createClient();
        VehicleLicenseOCRRequest request = new VehicleLicenseOCRRequest();
        request.setImageBase64(base64Img);
        if (index == DriverLicense.FRONT.type()) {
            request.setCardSide("FRONT");
        } else if (index == DriverLicense.BACK.type()) {
            request.setCardSide("BACK");
        } else {
            throw new GlobalException(504, "无效的行驶证面类型: " + index);
        }
        return client.VehicleLicenseOCR(request);
    }
    private com.tencentcloudapi.ocr.v20181119.OcrClient createClient() throws TencentCloudSDKException {
        Credential cred = new Credential(tencentOcr.getId(), tencentOcr.getSecret());
        HttpProfile httpProfile = new HttpProfile();
        httpProfile.setEndpoint("ocr.tencentcloudapi.com");
        ClientProfile clientProfile = new ClientProfile();
        clientProfile.setHttpProfile(httpProfile);
        com.tencentcloudapi.ocr.v20181119.OcrClient client = new com.tencentcloudapi.ocr.v20181119.OcrClient(cred, "ap-guangzhou", clientProfile);
        return client;
    }
}

如需笔者的工具类等，请私信笔者。