xiajingrui/id_ocr
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

init

Browse Source
This commit is contained in:
xiajingrui
2024-01-17 15:18:16 +08:00
commit 1bd4060e68
1514 changed files with 193581 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
deploy/fastdeploy/rockchip/cpp/CMakeLists.txt
+14
View File
@@ -0,0 +1,14 @@
PROJECT(infer_demo C CXX)
CMAKE_MINIMUM_REQUIRED (VERSION 3.10)
# 指定下载解压后的fastdeploy库路径
option(FASTDEPLOY_INSTALL_DIR "Path of downloaded fastdeploy sdk.")
include(${FASTDEPLOY_INSTALL_DIR}/FastDeploy.cmake)
# 添加FastDeploy依赖头文件
include_directories(${FASTDEPLOY_INCS})
add_executable(infer_demo ${PROJECT_SOURCE_DIR}/infer.cc)
# 添加FastDeploy库依赖
target_link_libraries(infer_demo ${FASTDEPLOY_LIBS})
deploy/fastdeploy/rockchip/cpp/README.md
+128
View File
@@ -0,0 +1,128 @@
[English](README_CN.md) | 简体中文
# PP-OCRv3 RKNPU2 C++部署示例
本目录下提供`infer.cc`, 供用户完成PP-OCRv3在RKNPU2的部署.
## 1. 部署环境准备
在部署前，需确认以下两个步骤
- 1. 在部署前，需自行编译基于RKNPU2的预测库，参考文档[RKNPU2部署环境编译](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install#自行编译安装)
- 2. 同时请用户参考[FastDeploy RKNPU2资源导航](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/rknpu2.md)
## 2.部署模型准备
在部署前, 请准备好您所需要运行的推理模型, 您可以在[FastDeploy支持的PaddleOCR模型列表](../README.md)中下载所需模型.
同时, 在RKNPU2上部署PP-OCR系列模型时，我们需要把Paddle的推理模型转为RKNN模型.
由于rknn_toolkit2工具暂不支持直接从Paddle直接转换为RKNN模型，因此我们需要先将Paddle推理模型转为ONNX模型, 最后转为RKNN模型, 示例如下.
```bash
# 下载PP-OCRv3文字检测模型
wget https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar
tar -xvf ch_PP-OCRv3_det_infer.tar
# 下载文字方向分类器模型
wget https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
tar -xvf ch_ppocr_mobile_v2.0_cls_infer.tar
# 下载PP-OCRv3文字识别模型
wget https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar
tar -xvf ch_PP-OCRv3_rec_infer.tar
# 请用户自行安装最新发布版本的paddle2onnx, 转换模型到ONNX格式的模型
paddle2onnx --model_dir ch_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer.onnx \
--enable_dev_version True
paddle2onnx --model_dir ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ch_ppocr_mobile_v2.0_cls_infer/ch_ppocr_mobile_v2.0_cls_infer.onnx \
--enable_dev_version True
paddle2onnx --model_dir ch_PP-OCRv3_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ch_PP-OCRv3_rec_infer/ch_PP-OCRv3_rec_infer.onnx \
--enable_dev_version True
# 固定模型的输入shape
python -m paddle2onnx.optimize --input_model ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer.onnx \
--output_model ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer.onnx \
--input_shape_dict "{'x':[1,3,960,960]}"
python -m paddle2onnx.optimize --input_model ch_ppocr_mobile_v2.0_cls_infer/ch_ppocr_mobile_v2.0_cls_infer.onnx \
--output_model ch_ppocr_mobile_v2.0_cls_infer/ch_ppocr_mobile_v2.0_cls_infer.onnx \
--input_shape_dict "{'x':[1,3,48,192]}"
python -m paddle2onnx.optimize --input_model ch_PP-OCRv3_rec_infer/ch_PP-OCRv3_rec_infer.onnx \
--output_model ch_PP-OCRv3_rec_infer/ch_PP-OCRv3_rec_infer.onnx \
--input_shape_dict "{'x':[1,3,48,320]}"
# 在rockchip/rknpu2_tools/目录下, 我们为用户提供了转换ONNX模型到RKNN模型的工具
python rockchip/rknpu2_tools/export.py --config_path tools/rknpu2/config/ppocrv3_det.yaml \
--target_platform rk3588
python rockchip/rknpu2_tools/export.py --config_path tools/rknpu2/config/ppocrv3_rec.yaml \
--target_platform rk3588
python rockchip/rknpu2_tools/export.py --config_path tools/rknpu2/config/ppocrv3_cls.yaml \
--target_platform rk3588
```
## 3.运行部署示例
在本目录执行如下命令即可完成编译测试，支持此模型需保证FastDeploy版本1.0.3以上(x.x.x>1.0.3), RKNN版本在1.4.1b22以上。
```
# 下载部署示例代码
git clone https://github.com/PaddlePaddle/FastDeploy.git
cd FastDeploy/examples/vision/ocr/PP-OCR/rockchip/cpp
# 如果您希望从PaddleOCR下载示例代码，请运行
git clone https://github.com/PaddlePaddle/PaddleOCR.git
# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到dygraph分支
git checkout dygraph
cd PaddleOCR/deploy/fastdeploy/rockchip/cpp
mkdir build
cd build
# 使用编译完成的FastDeploy库编译infer_demo
cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-rockchip
make -j
# 下载图片和字典文件
wget https://gitee.com/paddlepaddle/PaddleOCR/raw/release/2.6/doc/imgs/12.jpg
wget https://gitee.com/paddlepaddle/PaddleOCR/raw/release/2.6/ppocr/utils/ppocr_keys_v1.txt
# 拷贝RKNN模型到build目录
# CPU推理
./infer_demo ./ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer.onnx \
./ch_ppocr_mobile_v2.0_cls_infer/ch_ppocr_mobile_v2.0_cls_infer.onnx \
./ch_PP-OCRv3_rec_infer/ch_PP-OCRv3_rec_infer.onnx \
./ppocr_keys_v1.txt \
./12.jpg \
0
# RKNPU推理
./infer_demo ./ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer_rk3588_unquantized.rknn \
./ch_ppocr_mobile_v2.0_cls_infer/ch_ppocr_mobile_v20_cls_infer_rk3588_unquantized.rknn \
./ch_PP-OCRv3_rec_infer/ch_PP-OCRv3_rec_infer_rk3588_unquantized.rknn \
./ppocr_keys_v1.txt \
./12.jpg \
1
```
运行完成可视化结果如下图所示:
<img width="640" src="https://user-images.githubusercontent.com/109218879/185826024-f7593a0c-1bd2-4a60-b76c-15588484fa08.jpg">
结果输出如下:
```text
det boxes: [[276,174],[285,173],[285,178],[276,179]]rec text: rec score:0.000000 cls label: 1 cls score: 0.766602
det boxes: [[43,408],[483,390],[483,431],[44,449]]rec text: 上海斯格威铂尔曼大酒店 rec score:0.888450 cls label: 0 cls score: 1.000000
det boxes: [[186,456],[399,448],[399,480],[186,488]]rec text: 打浦路15号 rec score:0.988769 cls label: 0 cls score: 1.000000
det boxes: [[18,501],[513,485],[514,537],[18,554]]rec text: 绿洲仕格维花园公寓 rec score:0.992730 cls label: 0 cls score: 1.000000
det boxes: [[78,553],[404,541],[404,573],[78,585]]rec text: 打浦路252935号 rec score:0.983545 cls label: 0 cls score: 1.000000
Visualized result saved in ./vis_result.jpg
```
## 4. 更多指南
- [PP-OCR系列 C++ API查阅](https://www.paddlepaddle.org.cn/fastdeploy-api-doc/cpp/html/namespacefastdeploy_1_1vision_1_1ocr.html)
- [FastDeploy部署PaddleOCR模型概览](../../)
- [PP-OCRv3 Python部署](../python)
- [FastDeploy RKNPU2资源导航](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/rknpu2.md)
- 如果用户想要调整前后处理超参数、单独使用文字检测识别模型、使用其他模型等，更多详细文档与说明请参考[PP-OCR系列在CPU/GPU上的部署](../../cpu-gpu/cpp/README.md)
deploy/fastdeploy/rockchip/cpp/infer.cc
+126
View File
@@ -0,0 +1,126 @@
// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "fastdeploy/vision.h"
void InitAndInfer(const std::string &det_model_file,
const std::string &cls_model_file,
const std::string &rec_model_file,
const std::string &rec_label_file,
const std::string &image_file,
const fastdeploy::RuntimeOption &option,
const fastdeploy::ModelFormat &format) {
auto det_params_file = "";
auto cls_params_file = "";
auto rec_params_file = "";
auto det_option = option;
auto cls_option = option;
auto rec_option = option;
if (format == fastdeploy::ONNX) {
std::cout << "ONNX Model" << std::endl;
}
auto det_model = fastdeploy::vision::ocr::DBDetector(
det_model_file, det_params_file, det_option, format);
auto cls_model = fastdeploy::vision::ocr::Classifier(
cls_model_file, cls_params_file, cls_option, format);
auto rec_model = fastdeploy::vision::ocr::Recognizer(
rec_model_file, rec_params_file, rec_label_file, rec_option, format);
if (format == fastdeploy::RKNN) {
cls_model.GetPreprocessor().DisableNormalize();
cls_model.GetPreprocessor().DisablePermute();
det_model.GetPreprocessor().DisableNormalize();
det_model.GetPreprocessor().DisablePermute();
rec_model.GetPreprocessor().DisableNormalize();
rec_model.GetPreprocessor().DisablePermute();
}
det_model.GetPreprocessor().SetStaticShapeInfer(true);
rec_model.GetPreprocessor().SetStaticShapeInfer(true);
assert(det_model.Initialized());
assert(cls_model.Initialized());
assert(rec_model.Initialized());
// The classification model is optional, so the PP-OCR can also be connected
// in series as follows auto ppocr_v3 =
// fastdeploy::pipeline::PPOCRv3(&det_model, &rec_model);
auto ppocr_v3 =
fastdeploy::pipeline::PPOCRv3(&det_model, &cls_model, &rec_model);
// When users enable static shape infer for rec model, the batch size of cls
// and rec model must to be set to 1.
ppocr_v3.SetClsBatchSize(1);
ppocr_v3.SetRecBatchSize(1);
if (!ppocr_v3.Initialized()) {
std::cerr << "Failed to initialize PP-OCR." << std::endl;
return;
}
auto im = cv::imread(image_file);
fastdeploy::vision::OCRResult result;
if (!ppocr_v3.Predict(im, &result)) {
std::cerr << "Failed to predict." << std::endl;
return;
}
std::cout << result.Str() << std::endl;
auto vis_im = fastdeploy::vision::VisOcr(im, result);
cv::imwrite("vis_result.jpg", vis_im);
std::cout << "Visualized result saved in ./vis_result.jpg" << std::endl;
}
int main(int argc, char *argv[]) {
if (argc < 7) {
std::cout << "Usage: infer_demo path/to/det_model path/to/cls_model "
"path/to/rec_model path/to/rec_label_file path/to/image "
"run_option, "
"e.g ./infer_demo ./ch_PP-OCRv3_det_infer "
"./ch_ppocr_mobile_v2.0_cls_infer ./ch_PP-OCRv3_rec_infer "
"./ppocr_keys_v1.txt ./12.jpg 0"
<< std::endl;
std::cout << "The data type of run_option is int, 0: run with cpu; 1: run "
"with ascend."
<< std::endl;
return -1;
}
fastdeploy::RuntimeOption option;
fastdeploy::ModelFormat format;
int flag = std::atoi(argv[6]);
if (flag == 0) {
option.UseCpu();
format = fastdeploy::ONNX;
} else if (flag == 1) {
option.UseRKNPU2();
format = fastdeploy::RKNN;
}
std::string det_model_dir = argv[1];
std::string cls_model_dir = argv[2];
std::string rec_model_dir = argv[3];
std::string rec_label_file = argv[4];
std::string test_image = argv[5];
InitAndInfer(det_model_dir, cls_model_dir, rec_model_dir, rec_label_file,
test_image, option, format);
return 0;
}