Skip to content

GitLab

Commit 6878b65f authored by shizhm's avatar shizhm
Browse files

完善代码和文档

parent 97c70f1a
Hide whitespace changes
Inline Side-by-side
Showing with 58 additions and 98 deletions
CMakeLists.txt
View file @ 6878b65f
......@@ -2,7 +2,7 @@
cmake_minimum_required(VERSION 3.5)
# 设置项目名
project(MIGraphX_Samples)
project(PaddleOCR_VLPR)
# 设置编译器
set(CMAKE_CXX_COMPILER g++)
......@@ -37,7 +37,7 @@ link_libraries(${LIBRARY})
# 添加源文件
set(SOURCE_FILES ${CMAKE_CURRENT_SOURCE_DIR}/Src/main.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/Sample.cpp
# ${CMAKE_CURRENT_SOURCE_DIR}/Src/Sample.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/PaddleOCR/VLPR/clipper.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/PaddleOCR/VLPR/OcrDB.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/PaddleOCR/VLPR/utility.cpp
......@@ -47,4 +47,4 @@ set(SOURCE_FILES ${CMAKE_CURRENT_SOURCE_DIR}/Src/main.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/Utility/Filesystem.cpp)
# 添加可执行目标
add_executable(MIGraphX_Samples ${SOURCE_FILES})
add_executable(PaddleOCR_VLPR ${SOURCE_FILES})
Python/PaddleOCR_infer_migraphx.py
View file @ 6878b65f
......@@ -271,7 +271,7 @@ class process_pred(object):
else:
conf_list.append(1)
text = ''.join(char_list)
result_list.append((text, round(np.mean(conf_list), 2)))
result_list.append(text)
return result_list
def __call__(self, preds, label=None):
......@@ -533,7 +533,8 @@ if __name__=='__main__':
start = time.time()
ocr_sys = det_rec_functions(image)
dt_boxes = ocr_sys.get_boxes()
results_info = ocr_sys.recognition_img(dt_boxes)
results = ocr_sys.recognition_img(dt_boxes)
results_info = results[0][0]
print('net forward time: {:.4f}'.format(time.time() - start))
print("############# OCR Results #############")
......
README.md
View file @ 6878b65f
......@@ -8,7 +8,9 @@
DBnet是一种基于分割的文本检测方法，相比传统分割方法需要设定固定阈值，该模型将二值化操作插入到分割网络中进行联合优化，通过网络学习可以自适应的预测图像中每一个像素点的阈值，能够在像素水平很好的检测自然场景下不同形状的文字。SVTR是一种端到端的文本识别模型，通过单个视觉模型就可以一站式解决特征提取和文本转录两个任务，同时也保证了更快的推理速度。百度PaddleOCR开源项目提供了车牌识别的预训练模型，本示例使用PaddleOCR提供的蓝绿黄牌识别模型进行推理。
## 构建安装
## C++版本推理
### 拉取镜像
在光源可拉取推理的docker镜像，PaddleOCR工程推荐的镜像如下：
......@@ -19,7 +21,7 @@ docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort1.14.0_migraphx3.0
### 安装Opencv依赖
```python
cd <path_to_migraphx_samples>
cd <path_to_paddleocr_migraphx>
sh ./3rdParty/InstallOpenCVDependences.sh
```
......@@ -59,9 +61,7 @@ export LD_LIBRARY_PATH=<path_to_migraphx_samples>/depend/lib/:$LD_LIBRARY_PATH
source ~/.bashrc
```
## 推理
### C++版本推理
### 运行示例
成功编译PaddleOCR车牌识别工程后，在build目录下输入如下命令运行该示例：
......@@ -70,18 +70,22 @@ source ~/.bashrc
export MIGRAPHX_DYNAMIC_SHAPE=1
# 运行示例
./MIGraphX 0
./PaddleOCR_VLPR
```
PaddleOCR车牌识别结果为：
```
[皖AD19906, 0.999067]
皖AD19906
```
### python版本推理
## python版本推理
PaddleOCR工程同时提供了python推理示例。
### 环境配置
PaddleOCR工程同时提供了python推理示例，可使用如下命令运行：
python推理示例推荐使用C++版本示例同一镜像，python依赖安装：
```
# 进入python示例目录
......@@ -89,15 +93,20 @@ cd ./Python
# 安装依赖
pip install -r requirements.txt
```
### 运行示例
# 运行程序
环境配置成功之后，可在当前目录执行如下指令运行程序推理：
```
python PaddleOCR_infer_migraphx.py
```
PaddleOCR车牌识别结果为：
```
[[('皖AD19906', 0.99)]]
皖AD19906
```
## 历史版本
......
Src/Sample.cpp deleted 100644 → 0
View file @ 97c70f1a
#include <Sample.h>
#include <opencv2/dnn.hpp>
#include <SimpleLog.h>
#include <Filesystem.h>
#include <VLPR.h>
#include <fstream>
using namespace std;
using namespace cv;
using namespace cv::dnn;
using namespace migraphx;
using namespace migraphxSamples;
void Sample_VLPR()
{
// 创建PaddleOCR车牌识别
VLPR vlpr;
InitializationParameterOfDB initParamOfDB;
InitializationParameterOfSVTR initParamOfSVTR;
vlpr.Initialize(initParamOfDB, initParamOfSVTR);
// 读取测试图片
cv:: Mat Image=cv::imread("../Resource/Images/vlpr.jpg", 1);
// 推理
std::vector<std::string> recTexts;
std::vector<float> recTextScores;
double time1 = getTickCount();
vlpr.Infer(Image, recTexts, recTextScores);
double time2 = getTickCount();
double elapsedTime = (time2 - time1)*1000 / getTickFrequency();
LOG_INFO(stdout, "inference time:%f ms\n", elapsedTime);
// 打印结果
for (int i = 0; i < recTexts.size(); i++)
{
printf("VLPR Result:[%s, %f]\n", recTexts[i].c_str(), recTextScores[i]);
}
}
\ No newline at end of file
Src/Sample.h deleted 100644 → 0
View file @ 97c70f1a
// 示例程序
#ifndef __SAMPLE_H__
#define __SAMPLE_H__
// VLPR sample
void Sample_VLPR();
#endif
\ No newline at end of file
Src/main.cpp
View file @ 6878b65f
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <Sample.h>
#include <opencv2/dnn.hpp>
#include <SimpleLog.h>
#include <Filesystem.h>
#include <VLPR.h>
#include <fstream>
using namespace std;
using namespace cv;
using namespace migraphx;
using namespace migraphxSamples;
void MIGraphXSamplesUsage(char* programName)
{
printf("Usage : %s <index> \n", programName);
printf("index:\n");
printf("\t 0) PaddleOCR VLPR sample.\n");
}
int main(int argc, char *argv[])
int main()
{
if (argc < 2 || argc > 2)
{
MIGraphXSamplesUsage(argv[0]);
return -1;
}
if (!strncmp(argv[1], "-h", 2))
// 创建PaddleOCR车牌识别
VLPR vlpr;
InitializationParameterOfDB initParamOfDB;
InitializationParameterOfSVTR initParamOfSVTR;
vlpr.Initialize(initParamOfDB, initParamOfSVTR);
// 读取测试图片
cv:: Mat Image=cv::imread("../Resource/Images/2.jpg", 1);
// 推理
std::vector<std::string> recTexts;
std::vector<float> recTextScores;
double time1 = getTickCount();
vlpr.Infer(Image, recTexts, recTextScores);
double time2 = getTickCount();
double elapsedTime = (time2 - time1)*1000 / getTickFrequency();
LOG_INFO(stdout, "inference time:%f ms\n", elapsedTime);
// 打印结果
for (int i = 0; i < recTexts.size(); i++)
{
MIGraphXSamplesUsage(argv[0]);
return 0;
}
switch (*argv[1])
{
case '0':
{
Sample_VLPR();
break;
}
default :
{
MIGraphXSamplesUsage(argv[0]);
break;
}
printf("VLPR Result:%s\n", recTexts[i].c_str());
}
return 0;
}
\ No newline at end of file
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment