iOS Frameworks如何配置模型文件地址

[zcs110]

// 配置车牌识别参数
configuration.models_path = model_path; // 模型文件夹路径
这个model_path在iOS Frameworks中如何配置呢。

[kennenfromchina]

同问，iOS编译好framework之后，如何使用呢？参照linux代码去使用要初始化，model_path怎么配置？

[kennenfromchina]

我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了，但是有一个问题，我编译好的framework直接在项目中使用是可以的，但是打包成一个podsepc给Flutter用就编译失败，提示/chinese_license_plate_recognition/example/ios/Pods/Headers/Public/chinese_license_plate_recognition/chinese_license_plate_recognition-umbrella.h:14:9 'hyper_lpr_sdk.h' file not found
以下是我的podspec文件

#
# To learn more about a Podspec see http://guides.cocoapods.org/syntax/podspec.html.
# Run `pod lib lint chinese_license_plate_recognition.podspec` to validate before publishing.
#
Pod::Spec.new do |s|
  s.name             = 'chinese_license_plate_recognition'
  s.version          = '0.0.1'
  s.summary          = 'A new Flutter plugin project.'
  s.description      = <<-DESC
A new Flutter plugin project.
                       DESC
  s.homepage         = 'http://example.com'
  s.license          = { :file => '../LICENSE' }
  s.author           = { 'Your Company' => '[email protected]' }
  s.source           = { :path => '.' }
  s.source_files = 'Classes/**/*'
  s.public_header_files = 'Classes/**/*.h'
  s.dependency 'Flutter'
  s.platform = :ios, '9.0'

  # 添加 CoreML.framework 作为依赖项
  s.frameworks = 'CoreML'

  # 指定 framework 目录
  s.vendored_frameworks = 'Classes/Frameworks/*.framework'
  
  s.exclude_files = "Classes/Frameworks/*.framework/*.plist"

  # 指定 models 目录
  s.resource_bundles = {
    'chinese_license_plate_recognition' => ['Classes/models/*.mnn']
  }

  # Flutter.framework does not contain a i386 slice.
  s.pod_target_xcconfig = { 'DEFINES_MODULE' => 'YES', 'EXCLUDED_ARCHS[sdk=iphonesimulator*]' => 'i386' }
end

项目目录为：

- ios
    - chinese_license_plate_recognition.podspec
    - Classes
      - Frameworks
        - hyperlpr3.framework
        - MNN.framework
        - opencv2.framework
      - models
        - b320_backbone_h.mnn
        - b320_header_h.mnn
        - b640x_backbone_h.mnn
        - b640x_head_h.mnn
        - litemodel_cls_96xh.mnn
        - rpv3_mdict_160h.mnn
      - ChineseLicensePlateRecognitionPlugin.h
      - ChineseLicensePlateRecognitionPlugin.m

[zcs110]

我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了
我配置好了环境 但是检测的时候报错
det_arch.cpp
cv::resize(bgr, resized_img, cv::Size(w, h));
[InferenceHelperMnn][180] ios use cpu backend.Reshape error: 480 -> 0
你可以发一个demo给我吗[email protected]

[kennenfromchina]

我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了
我配置好了环境 但是检测的时候报错
det_arch.cpp
cv::resize(bgr, resized_img, cv::Size(w, h));
[InferenceHelperMnn][180] ios use cpu backend.Reshape error: 480 -> 0
你可以发一个demo给我吗[email protected]

https://github.com/kennenfromchina/LPRDemo

[kennenfromchina]

双层车牌识别请问怎么配置？

[qianxiaoer]

哈罗，邮件收到了，我会尽快查看，回复，谢谢！

[wisdom1314]

我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了
我配置好了环境 但是检测的时候报错
det_arch.cpp
cv::resize(bgr, resized_img, cv::Size(w, h));
[InferenceHelperMnn][180] ios use cpu backend.Reshape error: 480 -> 0
你可以发一个demo给我吗[email protected]

https://github.com/kennenfromchina/LPRDemo

你好，你的这个demo我跑通去扫车牌。识别率很低。是还需要配置什么吗

[kennenfromchina]

试试这样配置 // Configure license plate recognition parameters HLPR_ContextConfiguration configuration = {0}; configuration.models_path = (char *) modelPathC; configuration.max_num = 1; configuration.det_level = DETECT_LEVEL_HIGH; configuration.use_half = false; configuration.nms_threshold = 0.85f; configuration.rec_confidence_threshold = 0.95f; configuration.box_conf_threshold = 0.50f; configuration.threads = 1;

…

2024年12月12日 15:29，wisdomZhang ***@***.***> 写道： 我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了 我配置好了环境 但是检测的时候报错 det_arch.cpp cv::resize(bgr, resized_img, cv::Size(w, h)); [InferenceHelperMnn][180] ios use cpu backend.Reshape error: 480 -> 0 ***@***.*** ***@***.***> https://github.com/kennenfromchina/LPRDemo 你好，你的这个demo我跑通去扫车牌。识别率很低。是还需要配置什么吗 — Reply to this email directly, view it on GitHub <#388 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADRLTLHQ2DKBQHMB2WFD2PL2FE3NHAVCNFSM6AAAAABRMX7W5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMZYGAYTINZRGI>. You are receiving this because you commented.

[wisdom1314]

试试这样配置 // Configure license plate recognition parameters HLPR_ContextConfiguration configuration = {0}; configuration.models_path = (char ) modelPathC; configuration.max_num = 1; configuration.det_level = DETECT_LEVEL_HIGH; configuration.use_half = false; configuration.nms_threshold = 0.85f; configuration.rec_confidence_threshold = 0.95f; configuration.box_conf_threshold = 0.50f; configuration.threads = 1;
…
2024年12月12日 15:29，wisdomZhang @.> 写道： 我配置好了，把models/r2_mobile里面的文件拷贝到项目目录，models_path设置为项目根目录就行了 我配置好了环境 但是检测的时候报错 det_arch.cpp cv::resize(bgr, resized_img, cv::Size(w, h)); [InferenceHelperMnn][180] ios use cpu backend.Reshape error: 480 -> 0 @.* @.***> https://github.com/kennenfromchina/LPRDemo 你好，你的这个demo我跑通去扫车牌。识别率很低。是还需要配置什么吗 — Reply to this email directly, view it on GitHub <#388 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADRLTLHQ2DKBQHMB2WFD2PL2FE3NHAVCNFSM6AAAAABRMX7W5OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMZYGAYTINZRGI. You are receiving this because you commented.

好的，谢谢