PowLu/TencentOS-tiny
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

完成tflite_micro库目录结构，完成496TF移植部分文档

Browse Source
This commit is contained in:
QingChuanWS
2020-12-23 10:29:25 +08:00
parent 19280a8a79
commit c8eef23602
811 changed files with 219052 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/TFlite_Micro_Demo移植参考指南（Keil版）.md
View File

@@ -60,16 +60,16 @@
以下是整个例程的目录规划：
| 一级目录 | 二级目录 | 三级目录 | 说明 |
| :-------: | :--------------------------: | :------: | :----------------------------------------------------------: |
| arch | arm | | TencentOS tiny适配的IP核架构（含M核中断、调度、tick相关代码） |
| board | NUCLEO_STM32L496ZG | | 移植目标芯片的工程文件 |
| | | BSP | 板级支持包，外设驱动代码在Hardware目录 |
| component | tflite_micro | | tflite_micro源码 |
| examples | tflitemicro_person_detection | | 行人检测demo示例 |
| kernel | core | | TencentOS tiny内核源码 |
| | pm | | TencentOS tiny低功耗模块源码 |
| osal | cmsis_os | | TencentOS tiny提供的cmsis os 适配 |
| 一级目录 | 二级目录 | 三级目录 | 说明 |
| :-------: | :--------------------------: | :----------: | :----------------------------------------------------------: |
| arch | arm | | TencentOS tiny适配的IP核架构（含M核中断、调度、tick相关代码） |
| board | NUCLEO_STM32L496ZG | | 移植目标芯片的工程文件 |
| | | BSP | 板级支持包，外设驱动代码在Hardware目录 |
| component | ai | tflite_micro | tflite_micro源码 |
| examples | tflitemicro_person_detection | | 行人检测demo示例 |
| kernel | core | | TencentOS tiny内核源码 |
| | pm | | TencentOS tiny低功耗模块源码 |
| osal | cmsis_os | | TencentOS tiny提供的cmsis os 适配 |
完成TencentOS tiny基础keil工程准备工作后，在这个keil工程的基础上继续添加外设驱动代码。
@@ -128,23 +128,59 @@ void HAL_DCMI_FrameEventCallback(DCMI_HandleTypeDef *hdcmi)
## 三、Tensorflow Lite Micro移植
### 1. tflite_micro源码加入到keil工程
### 1. tflite_micro组件加入到keil工程
由于NUCLEO-L496ZG芯片中的内核为ARM Cortex M4，所以本次我们可以直接使用ARM Cortex M4版本的tensorflow_lite_micro.lib库来简化tflite_micro搭建流程。
#### 1.1 在project中加入新的文件夹tensorflow：
<div align=center>
<img src="./image/tflu_tensorflow文件夹增加的内容.png" width=80% />
</div>
#### 1.2 添加本次与行人检测demo有关的源文件
<div align=center>
<img src="./image/tflu_需要添加的文件.png" width=80% />
</div>
### 2. 解决编译错误
其中，retarget.c的路径为：`TencentOS-tiny\components\ai\tflite_micro\KEIL\retarget.c`
tensorflow_lite_micro.lib的路径为：`TencentOS-tiny\components\ai\tflite_micro\ARM_CortexM4_lib\tensorflow_lite_micro.lib`
其余.cc文件均在当前目录下的`tflu_person_detection`文件夹中。
#### 1.3 关闭Keil的MicroLib库：
<div align=center>
<img src="./image/tflu_取消Microlib.png" width=80% />
</div>
#### 1.4 添加tflite_micro需要的头文件：
<div align=center>
<img src="./image/tflu_添加include.png" width=80% />
</div>
注：最下方的路径为：
```
TencentOS-tiny\components\ai\tflite_micro\ARM_CortexM4_lib\tensorflow\lite\micro\tools\make\downloads
```
#### 1.5 调整优化等级和tflite_micro的交互信息输出串口：
<div align=center>
<img src="./image/tflu_STM32496宏.png" width=80% />
</div>
其中宏`NUCLEO_STM32L496ZG`是指定Nucleo STM32L496的hlpuart1为系统printf函数的输出串口，具体定义在Nucleo STM32L496的BSP文件夹中的`mcu_init.c`中。
### 2. 编写Person_Detection 任务函数
### 3. 测试行人检测模型
### 4. 编写Person_Detection 任务函数
#### 4.1 图像预处理
#### 2.1 图像预处理
<div align=center>
<img src="./image/RGB565.jpg" width=50% />
@@ -174,9 +210,7 @@ void input_convert(uint16_t* camera_buffer , uint8_t* model_buffer)
}
```
#### 4.2 行人检测线程任务函数
#### 3.2 行人检测线程任务函数
```c
void task1(void *arg)

76
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/TencentOS_tiny.uvoptx
View File

@@ -12,7 +12,7 @@
<lExt>*.lib</lExt>
<tExt>*.txt; *.h; *.inc</tExt>
<pExt>*.plm</pExt>
<CppX>*.cpp</CppX>
<CppX>*.cpp;*.cc</CppX>
<nMigrate>0</nMigrate>
</Extensions>
@@ -1051,7 +1051,7 @@
<Group>
<GroupName>hal</GroupName>
<tvExp>1</tvExp>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<cbSel>0</cbSel>
<RteFlg>0</RteFlg>
@@ -1119,7 +1119,7 @@
<Group>
<GroupName>examples</GroupName>
<tvExp>1</tvExp>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<cbSel>0</cbSel>
<RteFlg>0</RteFlg>
@@ -1146,24 +1146,84 @@
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>71</FileNumber>
<FileType>1</FileType>
<FileType>8</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>..\..\..\..\components\tflite_micro\KEIL\regarget.c</PathWithFileName>
<FilenameWithoutPath>regarget.c</FilenameWithoutPath>
<PathWithFileName>.\tflu_person_detection\person_detect_model_data.cc</PathWithFileName>
<FilenameWithoutPath>person_detect_model_data.cc</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>72</FileNumber>
<FileType>8</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>.\tflu_person_detection\model_settings.cc</PathWithFileName>
<FilenameWithoutPath>model_settings.cc</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>73</FileNumber>
<FileType>8</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>.\tflu_person_detection\main_functions.cc</PathWithFileName>
<FilenameWithoutPath>main_functions.cc</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>74</FileNumber>
<FileType>8</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>.\tflu_person_detection\image_provider.cc</PathWithFileName>
<FilenameWithoutPath>image_provider.cc</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>75</FileNumber>
<FileType>8</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>.\tflu_person_detection\detection_responder.cc</PathWithFileName>
<FilenameWithoutPath>detection_responder.cc</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>76</FileNumber>
<FileType>1</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>..\..\..\..\components\ai\tflite_micro\KEIL\retarget.c</PathWithFileName>
<FilenameWithoutPath>retarget.c</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>
<File>
<GroupNumber>10</GroupNumber>
<FileNumber>77</FileNumber>
<FileType>4</FileType>
<tvExp>0</tvExp>
<tvExpOptDlg>0</tvExpOptDlg>
<bDave2>0</bDave2>
<PathWithFileName>..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\tflm_person_detection.lib</PathWithFileName>
<FilenameWithoutPath>tflm_person_detection.lib</FilenameWithoutPath>
<PathWithFileName>..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\tensorflow_lite_micro_M4.lib</PathWithFileName>
<FilenameWithoutPath>tensorflow_lite_micro_M4.lib</FilenameWithoutPath>
<RteFlg>0</RteFlg>
<bShared>0</bShared>
</File>

43
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/TencentOS_tiny.uvprojx
View File

@@ -16,8 +16,8 @@
<TargetCommonOption>
<Device>STM32L496ZGTx</Device>
<Vendor>STMicroelectronics</Vendor>
<PackID>Keil.STM32L4xx_DFP.2.4.0</PackID>
<PackURL>http://www.keil.com/pack/</PackURL>
<PackID>Keil.STM32L4xx_DFP.2.5.0</PackID>
<PackURL>https://www.keil.com/pack/</PackURL>
<Cpu>IRAM(0x20000000-0x2004FFFF) IROM(0x8000000-0x80FFFFF) CLOCK(8000000) FPU2 CPUTYPE("Cortex-M4")</Cpu>
<FlashUtilSpec></FlashUtilSpec>
<StartupFile></StartupFile>
@@ -313,7 +313,7 @@
</ArmAdsMisc>
<Cads>
<interw>1</interw>
<Optim>7</Optim>
<Optim>4</Optim>
<oTime>0</oTime>
<SplitLS>0</SplitLS>
<OneElfS>1</OneElfS>
@@ -339,7 +339,7 @@
<MiscControls></MiscControls>
<Define>USE_HAL_DRIVER,STM32L496xx,NUCLEO_STM32L496ZG</Define>
<Undefine></Undefine>
<IncludePath>..\..\BSP\Inc;..\..\..\..\platform\vendor_bsp\st\STM32L4xx_HAL_Driver\Inc;..\..\..\..\platform\vendor_bsp\st\STM32L4xx_HAL_Driver\Inc\Legacy;..\..\..\..\platform\vendor_bsp\st\CMSIS\Device\ST\STM32L4xx\Include;..\..\..\..\platform\vendor_bsp\st\CMSIS\Include;..\..\..\..\arch\arm\arm-v7m\common\include;..\..\..\..\arch\arm\arm-v7m\cortex-m4\armcc;..\..\..\..\kernel\core\include;..\..\..\..\kernel\pm\include;..\..\..\..\osal\cmsis_os;..\..\..\..\examples\hello_world;..\..\TOS_CONFIG;..\..\..\..\net\at\include;..\..\..\..\kernel\hal\include;..\..\BSP\Hardware\Inc;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\third_party\flatbuffers\include;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\third_party\gemmlowp;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\third_party\kissfft;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\third_party\ruy;..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\tensorflow\lite\micro\tools\make\downloads</IncludePath>
<IncludePath>..\..\BSP\Inc;..\..\..\..\platform\vendor_bsp\st\STM32L4xx_HAL_Driver\Inc;..\..\..\..\platform\vendor_bsp\st\STM32L4xx_HAL_Driver\Inc\Legacy;..\..\..\..\platform\vendor_bsp\st\CMSIS\Device\ST\STM32L4xx\Include;..\..\..\..\platform\vendor_bsp\st\CMSIS\Include;..\..\..\..\arch\arm\arm-v7m\common\include;..\..\..\..\arch\arm\arm-v7m\cortex-m4\armcc;..\..\..\..\kernel\core\include;..\..\..\..\kernel\pm\include;..\..\..\..\osal\cmsis_os;..\..\..\..\examples\hello_world;..\..\TOS_CONFIG;..\..\..\..\net\at\include;..\..\..\..\kernel\hal\include;..\..\BSP\Hardware\Inc;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\third_party\flatbuffers\include;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\third_party\gemmlowp;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\third_party\kissfft;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\third_party\ruy;..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\tensorflow\lite\micro\tools\make\downloads</IncludePath>
</VariousControls>
</Cads>
<Aads>
@@ -779,14 +779,39 @@
<GroupName>tensorflow</GroupName>
<Files>
<File>
<FileName>regarget.c</FileName>
<FileType>1</FileType>
<FilePath>..\..\..\..\components\tflite_micro\KEIL\regarget.c</FilePath>
<FileName>person_detect_model_data.cc</FileName>
<FileType>8</FileType>
<FilePath>.\tflu_person_detection\person_detect_model_data.cc</FilePath>
</File>
<File>
<FileName>tflm_person_detection.lib</FileName>
<FileName>model_settings.cc</FileName>
<FileType>8</FileType>
<FilePath>.\tflu_person_detection\model_settings.cc</FilePath>
</File>
<File>
<FileName>main_functions.cc</FileName>
<FileType>8</FileType>
<FilePath>.\tflu_person_detection\main_functions.cc</FilePath>
</File>
<File>
<FileName>image_provider.cc</FileName>
<FileType>8</FileType>
<FilePath>.\tflu_person_detection\image_provider.cc</FilePath>
</File>
<File>
<FileName>detection_responder.cc</FileName>
<FileType>8</FileType>
<FilePath>.\tflu_person_detection\detection_responder.cc</FilePath>
</File>
<File>
<FileName>retarget.c</FileName>
<FileType>1</FileType>
<FilePath>..\..\..\..\components\ai\tflite_micro\KEIL\retarget.c</FilePath>
</File>
<File>
<FileName>tensorflow_lite_micro_M4.lib</FileName>
<FileType>4</FileType>
<FilePath>..\..\..\..\components\tflite_micro\ARM_CortexM4_lib\tflm_person_detection.lib</FilePath>
<FilePath>..\..\..\..\components\ai\tflite_micro\ARM_CortexM4_lib\tensorflow_lite_micro_M4.lib</FilePath>
</File>
</Files>
</Group>

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/NN Compiler.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 535 KiB

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/tflu_STM32496宏.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 139 KiB

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/tflu_tensorflow文件夹增加的内容.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 67 KiB

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/tflu_取消Microlib.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 113 KiB

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/tflu_添加include.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 170 KiB

BIN
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/image/tflu_需要添加的文件.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 82 KiB

25
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/detection_responder.cc Normal file
View File

@@ -0,0 +1,25 @@
/* Copyright 2019 The TensorFlow 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 "tensorflow/lite/micro/examples/person_detection_experimental/detection_responder.h"
// This dummy implementation writes person and no person scores to the error
// console. Real applications will want to take some custom action instead, and
// should implement their own versions of this function.
void RespondToDetection(tflite::ErrorReporter* error_reporter,
int8_t person_score, int8_t no_person_score) {
TF_LITE_REPORT_ERROR(error_reporter, "person score:%d no person score %d",
person_score, no_person_score);
}

34
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/detection_responder.h Normal file
View File

@@ -0,0 +1,34 @@
/* Copyright 2019 The TensorFlow 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.
==============================================================================*/
// Provides an interface to take an action based on the output from the person
// detection model.
#ifndef TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_DETECTION_RESPONDER_H_
#define TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_DETECTION_RESPONDER_H_
#include "tensorflow/lite/c/common.h"
#include "tensorflow/lite/micro/micro_error_reporter.h"
// Called every time the results of a person detection run are available. The
// `person_score` has the numerical confidence that the captured image contains
// a person, and `no_person_score` has the numerical confidence that the image
// does not contain a person. Typically if person_score > no person score, the
// image is considered to contain a person. This threshold may be adjusted for
// particular applications.
void RespondToDetection(tflite::ErrorReporter* error_reporter,
int8_t person_score, int8_t no_person_score);
#endif // TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_DETECTION_RESPONDER_H_

26
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/image_provider.cc Normal file
View File

@@ -0,0 +1,26 @@
/* Copyright 2019 The TensorFlow 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 "tensorflow/lite/micro/examples/person_detection_experimental/image_provider.h"
#include "tensorflow/lite/micro/examples/person_detection_experimental/model_settings.h"
TfLiteStatus GetImage(tflite::ErrorReporter* error_reporter, int image_width,
int image_height, int channels, int8_t* image_data,
uint8_t * hardware_input) {
for (int i = 0; i < image_width * image_height * channels; ++i) {
image_data[i] = hardware_input[i];
}
return kTfLiteOk;
}

40
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/image_provider.h Normal file
View File

@@ -0,0 +1,40 @@
/* Copyright 2019 The TensorFlow 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.
==============================================================================*/
#ifndef TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_IMAGE_PROVIDER_H_
#define TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_IMAGE_PROVIDER_H_
#include "tensorflow/lite/c/common.h"
#include "tensorflow/lite/micro/micro_error_reporter.h"
// This is an abstraction around an image source like a camera, and is
// expected to return 8-bit sample data. The assumption is that this will be
// called in a low duty-cycle fashion in a low-power application. In these
// cases, the imaging sensor need not be run in a streaming mode, but rather can
// be idled in a relatively low-power mode between calls to GetImage(). The
// assumption is that the overhead and time of bringing the low-power sensor out
// of this standby mode is commensurate with the expected duty cycle of the
// application. The underlying sensor may actually be put into a streaming
// configuration, but the image buffer provided to GetImage should not be
// overwritten by the driver code until the next call to GetImage();
//
// The reference implementation can have no platform-specific dependencies, so
// it just returns a static image. For real applications, you should
// ensure there's a specialized implementation that accesses hardware APIs.
TfLiteStatus GetImage(tflite::ErrorReporter* error_reporter, int image_width,
int image_height, int channels, int8_t* image_data,
uint8_t * hardware_input);
#endif // TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_IMAGE_PROVIDER_H_

119
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/main_functions.cc Normal file
View File

@@ -0,0 +1,119 @@
/* Copyright 2019 The TensorFlow 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 "tensorflow/lite/micro/examples/person_detection_experimental/main_functions.h"
#include "tensorflow/lite/micro/examples/person_detection_experimental/detection_responder.h"
#include "tensorflow/lite/micro/examples/person_detection_experimental/image_provider.h"
#include "tensorflow/lite/micro/examples/person_detection_experimental/model_settings.h"
#include "tensorflow/lite/micro/examples/person_detection_experimental/person_detect_model_data.h"
#include "tensorflow/lite/micro/micro_error_reporter.h"
#include "tensorflow/lite/micro/micro_interpreter.h"
#include "tensorflow/lite/micro/micro_mutable_op_resolver.h"
#include "tensorflow/lite/schema/schema_generated.h"
#include "tensorflow/lite/version.h"
// Globals, used for compatibility with Arduino-style sketches.
namespace {
tflite::ErrorReporter* error_reporter = nullptr;
const tflite::Model* model = nullptr;
tflite::MicroInterpreter* interpreter = nullptr;
TfLiteTensor* input = nullptr;
// In order to use optimized tensorflow lite kernels, a signed int8_t quantized
// model is preferred over the legacy unsigned model format. This means that
// throughout this project, input images must be converted from unisgned to
// signed format. The easiest and quickest way to convert from unsigned to
// signed 8-bit integers is to subtract 128 from the unsigned value to get a
// signed value.
// An area of memory to use for input, output, and intermediate arrays.
constexpr int kTensorArenaSize = 115 * 1024;
static uint8_t tensor_arena[kTensorArenaSize];
} // namespace
// The name of this function is important for Arduino compatibility.
void person_detect_init() {
// Set up logging. Google style is to avoid globals or statics because of
// lifetime uncertainty, but since this has a trivial destructor it's okay.
// NOLINTNEXTLINE(runtime-global-variables)
static tflite::MicroErrorReporter micro_error_reporter;
error_reporter = &micro_error_reporter;
// Map the model into a usable data structure. This doesn't involve any
// copying or parsing, it's a very lightweight operation.
model = tflite::GetModel(g_person_detect_model_data);
if (model->version() != TFLITE_SCHEMA_VERSION) {
TF_LITE_REPORT_ERROR(error_reporter,
"Model provided is schema version %d not equal "
"to supported version %d.",
model->version(), TFLITE_SCHEMA_VERSION);
return;
}
// Pull in only the operation implementations we need.
// This relies on a complete list of all the ops needed by this graph.
// An easier approach is to just use the AllOpsResolver, but this will
// incur some penalty in code space for op implementations that are not
// needed by this graph.
//
// tflite::AllOpsResolver resolver;
// NOLINTNEXTLINE(runtime-global-variables)
static tflite::MicroMutableOpResolver<5> micro_op_resolver;
micro_op_resolver.AddAveragePool2D();
micro_op_resolver.AddConv2D();
micro_op_resolver.AddDepthwiseConv2D();
micro_op_resolver.AddReshape();
micro_op_resolver.AddSoftmax();
// Build an interpreter to run the model with.
// NOLINTNEXTLINE(runtime-global-variables)
static tflite::MicroInterpreter static_interpreter(
model, micro_op_resolver, tensor_arena, kTensorArenaSize, error_reporter);
interpreter = &static_interpreter;
// Allocate memory from the tensor_arena for the model's tensors.
TfLiteStatus allocate_status = interpreter->AllocateTensors();
if (allocate_status != kTfLiteOk) {
TF_LITE_REPORT_ERROR(error_reporter, "AllocateTensors() failed");
return;
}
// Get information about the memory area to use for the model's input.
input = interpreter->input(0);
}
// The name of this function is important for Arduino compatibility.
int person_detect(uint8_t * hardware_input) {
// Get image from provider.
if (kTfLiteOk != GetImage(error_reporter, kNumCols, kNumRows, kNumChannels,
input->data.int8, hardware_input)) {
TF_LITE_REPORT_ERROR(error_reporter, "Image capture failed.");
}
// Run the model on this input and make sure it succeeds.
if (kTfLiteOk != interpreter->Invoke()) {
TF_LITE_REPORT_ERROR(error_reporter, "Invoke failed.");
}
TfLiteTensor* output = interpreter->output(0);
// Process the inference results.
int8_t person_score = output->data.uint8[kPersonIndex];
int8_t no_person_score = output->data.uint8[kNotAPersonIndex];
RespondToDetection(error_reporter, person_score, no_person_score);
if(person_score >= no_person_score + 50) return 1;
else return 0;
}

30
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/main_functions.h Normal file
View File

@@ -0,0 +1,30 @@
/* Copyright 2019 The TensorFlow 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.
==============================================================================*/
#ifndef TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MAIN_FUNCTIONS_H_
#define TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MAIN_FUNCTIONS_H_
#include "tensorflow/lite/c/common.h"
// Initializes all data needed for the example. The name is important, and needs
// to be setup() for Arduino compatibility.
extern "C" void person_detect_init();
// Runs one iteration of data gathering and inference. This should be called
// repeatedly from the application code. The name needs to be loop() for Arduino
// compatibility.
extern "C" int person_detect(uint8_t * hardware_input);
#endif // TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MAIN_FUNCTIONS_H_

21
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/model_settings.cc Normal file
View File

@@ -0,0 +1,21 @@
/* Copyright 2019 The TensorFlow 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 "tensorflow/lite/micro/examples/person_detection_experimental/model_settings.h"
const char* kCategoryLabels[kCategoryCount] = {
"notperson",
"person",
};

35
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/model_settings.h Normal file
View File

@@ -0,0 +1,35 @@
/* Copyright 2019 The TensorFlow 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.
==============================================================================*/
#ifndef TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MODEL_SETTINGS_H_
#define TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MODEL_SETTINGS_H_
// Keeping these as constant expressions allow us to allocate fixed-sized arrays
// on the stack for our working memory.
// All of these values are derived from the values used during model training,
// if you change your model you'll need to update these constants.
constexpr int kNumCols = 96;
constexpr int kNumRows = 96;
constexpr int kNumChannels = 1;
constexpr int kMaxImageSize = kNumCols * kNumRows * kNumChannels;
constexpr int kCategoryCount = 2;
constexpr int kPersonIndex = 1;
constexpr int kNotAPersonIndex = 0;
extern const char* kCategoryLabels[kCategoryCount];
#endif // TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_MODEL_SETTINGS_H_

23148
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/person_detect_model_data.cc Normal file
View File

File diff suppressed because it is too large Load Diff

27
board/NUCLEO_STM32L496ZG/KEIL/tflitemicro_person_detection/tflu_person_detection/person_detect_model_data.h Normal file
View File

@@ -0,0 +1,27 @@
/* Copyright 2019 The TensorFlow 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.
==============================================================================*/
// This is a standard TensorFlow Lite model file that has been converted into a
// C data array, so it can be easily compiled into a binary for devices that
// don't have a file system. It was created using the command:
// xxd -i person_detect.tflite > person_detect_model_data.cc
#ifndef TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_PERSON_DETECT_MODEL_DATA_H_
#define TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_PERSON_DETECT_MODEL_DATA_H_
extern const unsigned char g_person_detect_model_data[];
extern const int g_person_detect_model_data_len;
#endif // TENSORFLOW_LITE_MICRO_EXAMPLES_PERSON_DETECTION_EXPERIMENTAL_PERSON_DETECT_MODEL_DATA_H_