在树莓派上目标检测模型只能部署SSD?能不能换成效果更好的YoloV3?那当然是——可以！
只是Paddle-Lite-Demo项目上仅提供了SSD模型的部署方法，换成YoloV3恐怕少不了一番周折……
本项目提供了完整的安全帽检测YoloV3模型在树莓派上的部署教程，旨在帮助读者少走弯路，都来试试吧?
- 安全帽佩戴检测模型训练与一键部署（PaddleX、HubServing）
- PaddleX、PP-Yolo：手把手教你训练、加密、部署目标检测模型
- PaddleLite树莓派从0到1：安全帽检测小车部署（一）
- 巡检告警机器人上线！PaddleLite安全帽检测小车部署（二）
- YoloV3检测模型在树莓派上的部署（PaddleX、PaddleLite）
- 注1：关于PaddleDetection如何进行迁移学习训练，并导出可在树莓派上部署的Lite模型，请查看前置项目PaddleLite树莓派从0到1：安全帽检测小车部署（一）
- 注2：关于如何在设备上自行编译PaddleLite，请查看前置项目YoloV3检测模型在树莓派上的部署（PaddleX、PaddleLite）
- 注3：本项目使用的是Raspberry Pi OS 64位操作系统，如与您使用的系统不同，请到https://github.com/PaddlePaddle/Paddle-Lite/releases上选择或自行编译对应的预测库，但是部署代码无需调整。
在项目挂载的数据集中，提供了基于yolov3_mobilenetv3训练的安全帽检测部署模型det_yolov3_mobilenetv3.nb
!ls data/data62679/det_yolov3_mobilenetv3.nbdata/data62679/det_yolov3_mobilenetv3.nb
git clone https://gitee.com/paddlepaddle/paddle-lite
git checkout v2.6.3如果您的树莓派也是64位系统，可下载项目挂载数据集里的预编译库；如果不是，请参考源码编译方法自行编译
# Paddle-Lite(ArmLinuxV8)FULL预编译库.zip
!ls data/data62679/Paddle-Lite*data/data62679/Paddle-Lite(ArmLinuxV8)FULL预编译库.zip
解压后paddle-lite目录结构如下：
pi@raspberrypi:~ $ ls paddle-lite/
build.lite.armlinux.armv8.gcc cmake CMakeLists.txt docs LICENSE lite metal mobile README_en.md README.md third-party third-party-05b862.tar.gz tools web其实，在Paddle-Lite项目中已经给出了yolov3部署的demo，可参考https://gitee.com/paddlepaddle/paddle-lite/blob/develop/lite/demo/cxx/yolov3_detection/yolov3_detection.cc
但是yolov3_detection.cc只给出了一个文件，缺少相关的文档介绍，这里试着解读下代码的重点。
这也是近来几次课程里老师传授的重点了，先看main()函数有什么功能，可以很快理解代码的意图。
int main(int argc, char** argv) {
if (argc < 3) {
std::cerr << "[ERROR] usage: " << argv[0] << " model_file image_path\n";
exit(1);
}
// 传入Lite模型文件
std::string model_file = argv[1];
// 对图片预测，传入图片文件
std::string img_path = argv[2];
// 执行预测
RunModel(model_file, img_path);
return 0;
}
所以，`main()`函数其实很简单，我们需要传入两个参数，模型文件路径和图片路径，然后执行预测。毫无疑问RunModel(model_file, img_path)的实现是这个文件的重点，这里增加了一些注释，重点要注意的是处理图片大小的两个内置类型变量
void RunModel(std::string model_file, std::string img_path) {
// 1. Set MobileConfig 读取模型文件
MobileConfig config;
config.set_model_from_file(model_file);
// 2. Create PaddlePredictor by MobileConfig 创建Predictor
std::shared_ptr<PaddlePredictor> predictor =
CreatePaddlePredictor<MobileConfig>(config);
// 设置模型输入，注意到608*608在树莓派3B上会耗尽资源，这里要进行调整
const int in_width = 608;
const int in_height = 608;
// 3. Prepare input data from image
// input 0
std::unique_ptr<Tensor> input_tensor0(std::move(predictor->GetInput(0)));
input_tensor0->Resize({1, 3, in_height, in_width});
auto* data0 = input_tensor0->mutable_data<float>();
// 使用OpenCV把图片读取出来
cv::Mat img = imread(img_path, cv::IMREAD_COLOR);
// 前处理部分，注意到传入了in_width和in_height
pre_process(img, in_width, in_height, data0);
// input1
std::unique_ptr<Tensor> input_tensor1(std::move(predictor->GetInput(1)));
input_tensor1->Resize({1, 2});
auto* data1 = input_tensor1->mutable_data<int>();
data1[0] = img.rows;
data1[1] = img.cols;
// 4. Run predictor 执行预测
predictor->Run();
// 5. Get output and post process 得到预测结果并进行后处理
std::unique_ptr<const Tensor> output_tensor(
std::move(predictor->GetOutput(0)));
auto* outptr = output_tensor->data<float>();
auto shape_out = output_tensor->shape();
int64_t cnt = 1;
for (auto& i : shape_out) {
cnt *= i;
}
// 后处理部分，将检测结果和原图画框
auto rec_out = detect_object(outptr, static_cast<int>(cnt / 6), 0.5f, img);
// 保存画框后的图片
std::string result_name =
img_path.substr(0, img_path.find(".")) + "_yolov3_detection_result.jpg";
cv::imwrite(result_name, img);
}前处理函数的逻辑：得到一张图片，resize一番！再做一些归一化处理，都是常规动作，所以最后的neon_mean_scale()这里也不赘述了。
void pre_process(const cv::Mat& img, int width, int height, float* data) {
cv::Mat rgb_img;
cv::cvtColor(img, rgb_img, cv::COLOR_BGR2RGB);
cv::resize(
rgb_img, rgb_img, cv::Size(width, height), 0.f, 0.f, cv::INTER_CUBIC);
cv::Mat imgf;
rgb_img.convertTo(imgf, CV_32FC3, 1 / 255.f);
std::vector<float> mean = {0.485f, 0.456f, 0.406f};
std::vector<float> scale = {0.229f, 0.224f, 0.225f};
const float* dimg = reinterpret_cast<const float*>(imgf.data);
neon_mean_scale(dimg, data, width * height, mean, scale);
}后处理函数内容较多，但我们可以思考下，如果对后处理函数要进行改动，会改哪里?一般就是检测到目标，报个警之类的吧?那这时候，只要重点关注能在哪里加入（或直接利用）if语句吧?
至于截图?RunModel()里面单独拎出来了不是，紧接在后处理函数后面。
std::vector<Object> detect_object(const float* data,
int count,
float thresh,
cv::Mat& image) { // NOLINT
if (data == nullptr) {
std::cerr << "[ERROR] data can not be nullptr\n";
exit(1);
}
std::vector<Object> rect_out;
for (int iw = 0; iw < count; iw++) {
int oriw = image.cols;
int orih = image.rows;
if (data[1] > thresh) {
Object obj;
int x = static_cast<int>(data[2]);
int y = static_cast<int>(data[3]);
int w = static_cast<int>(data[4] - data[2] + 1);
int h = static_cast<int>(data[5] - data[3] + 1);
cv::Rect rec_clip =
cv::Rect(x, y, w, h) & cv::Rect(0, 0, image.cols, image.rows);
obj.class_id = static_cast<int>(data[0]);
obj.prob = data[1];
obj.rec = rec_clip;
// 下面就是个很理想的，可以考虑直接加入报警函数的地方
if (w > 0 && h > 0 && obj.prob <= 1) {
rect_out.push_back(obj);
cv::rectangle(image, rec_clip, cv::Scalar(0, 0, 255), 1, cv::LINE_AA);
std::string str_prob = std::to_string(obj.prob);
std::string text = std::string(class_names[obj.class_id]) + ": " +
str_prob.substr(0, str_prob.find(".") + 4);
int font_face = cv::FONT_HERSHEY_COMPLEX_SMALL;
double font_scale = 1.f;
int thickness = 1;
cv::Size text_size =
cv::getTextSize(text, font_face, font_scale, thickness, nullptr);
float new_font_scale = w * 0.5 * font_scale / text_size.width;
text_size = cv::getTextSize(
text, font_face, new_font_scale, thickness, nullptr);
cv::Point origin;
origin.x = x + 3;
origin.y = y + text_size.height + 3;
cv::putText(image,
text,
origin,
font_face,
new_font_scale,
cv::Scalar(0, 255, 255),
thickness,
cv::LINE_AA);
std::cout << "detection, image size: " << image.cols << ", "
<< image.rows
<< ", detect object: " << class_names[obj.class_id]
<< ", score: " << obj.prob << ", location: x=" << x
<< ", y=" << y << ", width=" << w << ", height=" << h
<< std::endl;
}
}
data += 6;
}
return rect_out;
}理解了yolov3_detection.cc的结构，梳理下改动的思路
1. 修改resize后图片的大小（必须的，否则树莓派3B资源就爆了）
2. 写一个报警函数warn()——关于报警函数的写法和树莓派引脚接线，请查看巡检告警机器人上线！PaddleLite安全帽检测小车部署（二）
3. 利用后处理的if判断语句，增加检测到未佩戴安全帽时才触发报警
4. 截图存证，引入时间戳，这样可以连续保存函数
5. 将输入图片的预测转换为直接输入视频流
6. 使用cmake编译——关于CMakeLists.txt和build.sh的写法，请查看巡检告警机器人上线！PaddleLite安全帽检测小车部署（二）
篇幅所限，yolov3_detection.cc的修改就不赘述了，读者可以用前面的方法试着理解下。
// Copyright (c) 2019 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.
// 注意这里增加了wiringPi库和计算时间的sys/time库
#include <iostream>
#include <vector>
#include <sys/time.h>
#include <stdlib.h>
#include <string.h>
#include "opencv2/core.hpp"
#include "opencv2/imgcodecs.hpp"
#include "opencv2/imgproc.hpp"
#include <opencv2/opencv.hpp>
#include <opencv2/highgui.hpp>
#include <opencv2/core/core.hpp>
#include <wiringPi.h>
#include "paddle_api.h" // NOLINT
#define LED 17
using namespace paddle::lite_api; // NOLINT
struct Object {
cv::Rect rec;
int class_id;
float prob;
};
int64_t ShapeProduction(const shape_t& shape) {
int64_t res = 1;
for (auto i : shape) res *= i;
return res;
}
inline int64_t get_current_us() {
struct timeval time;
gettimeofday(&time, NULL);
return 1000000LL * (int64_t)time.tv_sec + (int64_t)time.tv_usec;
}
// 预测的标签是写死的，这里改成与训练数据集对应的标签
const char* class_names[] = { "head", "helmet", "person"
};
// 增加报警函数，注意使用的是BCM编码
void warn(void)
{
wiringPiSetupSys();
pinMode(LED, OUTPUT);
for (int i = 1; i <= 6; i++)
{
digitalWrite(LED, HIGH); // 启用
delay(500); // 毫秒
digitalWrite(LED, LOW); // 关
delay(500);
}
}
// fill tensor with mean and scale and trans layout: nhwc -> nchw, neon speed up
void neon_mean_scale(const float* din,
float* dout,
int size,
const std::vector<float> mean,
const std::vector<float> scale) {
if (mean.size() != 3 || scale.size() != 3) {
std::cerr << "[ERROR] mean or scale size must equal to 3\n";
exit(1);
}
float32x4_t vmean0 = vdupq_n_f32(mean[0]);
float32x4_t vmean1 = vdupq_n_f32(mean[1]);
float32x4_t vmean2 = vdupq_n_f32(mean[2]);
float32x4_t vscale0 = vdupq_n_f32(1.f / scale[0]);
float32x4_t vscale1 = vdupq_n_f32(1.f / scale[1]);
float32x4_t vscale2 = vdupq_n_f32(1.f / scale[2]);
float* dout_c0 = dout;
float* dout_c1 = dout + size;
float* dout_c2 = dout + size * 2;
int i = 0;
for (; i < size - 3; i += 4) {
float32x4x3_t vin3 = vld3q_f32(din);
float32x4_t vsub0 = vsubq_f32(vin3.val[0], vmean0);
float32x4_t vsub1 = vsubq_f32(vin3.val[1], vmean1);
float32x4_t vsub2 = vsubq_f32(vin3.val[2], vmean2);
float32x4_t vs0 = vmulq_f32(vsub0, vscale0);
float32x4_t vs1 = vmulq_f32(vsub1, vscale1);
float32x4_t vs2 = vmulq_f32(vsub2, vscale2);
vst1q_f32(dout_c0, vs0);
vst1q_f32(dout_c1, vs1);
vst1q_f32(dout_c2, vs2);
din += 12;
dout_c0 += 4;
dout_c1 += 4;
dout_c2 += 4;
}
for (; i < size; i++) {
*(dout_c0++) = (*(din++) - mean[0]) * scale[0];
*(dout_c0++) = (*(din++) - mean[1]) * scale[1];
*(dout_c0++) = (*(din++) - mean[2]) * scale[2];
}
}
// 前处理部分没有改动
void pre_process(const cv::Mat& img, int width, int height, float* data) {
cv::Mat rgb_img;
cv::cvtColor(img, rgb_img, cv::COLOR_BGR2RGB);
cv::resize(
rgb_img, rgb_img, cv::Size(width, height), 0.f, 0.f, cv::INTER_CUBIC);
cv::Mat imgf;
rgb_img.convertTo(imgf, CV_32FC3, 1 / 255.f);
std::vector<float> mean = { 0.485f, 0.456f, 0.406f };
std::vector<float> scale = { 0.229f, 0.224f, 0.225f };
const float* dimg = reinterpret_cast<const float*>(imgf.data);
neon_mean_scale(dimg, data, width * height, mean, scale);
}
// 后处理部分
std::vector<Object> detect_object(const float* data,
int count,
float thresh,
cv::Mat& image) { // NOLINT
if (data == nullptr) {
std::cerr << "[ERROR] data can not be nullptr\n";
exit(1);
}
std::vector<Object> rect_out;
for (int iw = 0; iw < count; iw++) {
int oriw = image.cols;
int orih = image.rows;
if (data[1] > thresh) {
Object obj;
int x = static_cast<int>(data[2]);
int y = static_cast<int>(data[3]);
int w = static_cast<int>(data[4] - data[2] + 1);
int h = static_cast<int>(data[5] - data[3] + 1);
cv::Rect rec_clip =
cv::Rect(x, y, w, h) & cv::Rect(0, 0, image.cols, image.rows);
obj.class_id = static_cast<int>(data[0]);
obj.prob = data[1];
obj.rec = rec_clip;
// 改为只检测到未佩戴安全帽，才开始画框和报警
if (w > 0 && h > 0 && obj.prob <= 1 && obj.class_id==0) {
// 新增的报警函数
warn();
rect_out.push_back(obj);
cv::rectangle(image, rec_clip, cv::Scalar(0, 0, 255), 1, cv::LINE_AA);
std::string str_prob = std::to_string(obj.prob);
std::string text = std::string(class_names[obj.class_id]) + ": " +
str_prob.substr(0, str_prob.find(".") + 4);
int font_face = cv::FONT_HERSHEY_COMPLEX_SMALL;
double font_scale = 1.f;
int thickness = 1;
cv::Size text_size =
cv::getTextSize(text, font_face, font_scale, thickness, nullptr);
float new_font_scale = w * 0.5 * font_scale / text_size.width;
text_size = cv::getTextSize(
text, font_face, new_font_scale, thickness, nullptr);
cv::Point origin;
origin.x = x + 3;
origin.y = y + text_size.height + 3;
cv::putText(image,
text,
origin,
font_face,
new_font_scale,
cv::Scalar(0, 255, 255),
thickness,
cv::LINE_AA);
std::cout << "detection, image size: " << image.cols << ", "
<< image.rows
<< ", detect object: " << class_names[obj.class_id]
<< ", score: " << obj.prob << ", location: x=" << x
<< ", y=" << y << ", width=" << w << ", height=" << h
<< std::endl;
}
}
data += 6;
}
return rect_out;
}
void RunModel(std::string model_file, const cv::Mat& img) {
// 1. Set MobileConfig
MobileConfig config;
config.set_model_from_file(model_file);
// 2. Create PaddlePredictor by MobileConfig
std::shared_ptr<PaddlePredictor> predictor =
CreatePaddlePredictor<MobileConfig>(config);
// 资源不足的话resize的图片一定要改小
const int in_width = 320;
const int in_height = 320;
// 3. Prepare input data from image
// input 0
std::unique_ptr<Tensor> input_tensor0(std::move(predictor->GetInput(0)));
input_tensor0->Resize({ 1, 3, in_height, in_width });
auto* data0 = input_tensor0->mutable_data<float>();
pre_process(img, in_width, in_height, data0);
// input1
std::unique_ptr<Tensor> input_tensor1(std::move(predictor->GetInput(1)));
input_tensor1->Resize({ 1, 2 });
auto* data1 = input_tensor1->mutable_data<int>();
data1[0] = img.rows;
data1[1] = img.cols;
//开始计时
char saveName[256];
int start_time = get_current_us();
// 4. Run predictor
predictor->Run();
//记录预测结束时间
int end_time = get_current_us();
//计算预测时间
double process_time = (end_time - start_time) / 1000.0f;
// 5. Get output and post process
std::unique_ptr<const Tensor> output_tensor(
std::move(predictor->GetOutput(0)));
auto* outptr = output_tensor->data<float>();
auto shape_out = output_tensor->shape();
int64_t cnt = 1;
for (auto& i : shape_out) {
cnt *= i;
}
//注意这里，要从摄像头里把图片复制一份（画框用）
cv::Mat output_image = img.clone();
auto rec_out = detect_object(outptr, static_cast<int>(cnt / 6), 0.5f, output_image);
// 把预测结束的时间戳作为文件名的一部分
std::string name = std::to_string(end_time);
std::string result_name = name + "_yolov3_detection_result.jpg";
// 截图存证
cv::imwrite(result_name, output_image);
printf("Preprocess time: %f ms\n", process_time);
}
int main(int argc, char** argv) {
// 只需传入模型文件
std::string model_file = argv[1];
// 纯摄像头调用
cv::VideoCapture cap(-1);
cap.set(CV_CAP_PROP_FRAME_WIDTH, 640);
cap.set(CV_CAP_PROP_FRAME_HEIGHT, 480);
if (!cap.isOpened()) {
return -1;
}
while (1) {
cv::Mat input_image;
cap >> input_image;
RunModel(model_file, input_image);
if (cv::waitKey(1) == char('q')) {
break;
}
}
cap.release();
cv::destroyAllWindows();
return 0;
}参考Paddle-Lite-Demo/PaddleLite-armlinux-demo/object_detection_demo
cmake_minimum_required(VERSION 3.10)
set(CMAKE_SYSTEM_NAME Linux)
if(TARGET_ARCH_ABI STREQUAL "armv8")
set(CMAKE_SYSTEM_PROCESSOR aarch64)
set(CMAKE_C_COMPILER "aarch64-linux-gnu-gcc")
set(CMAKE_CXX_COMPILER "aarch64-linux-gnu-g++")
elseif(TARGET_ARCH_ABI STREQUAL "armv7hf")
set(CMAKE_SYSTEM_PROCESSOR arm)
set(CMAKE_C_COMPILER "arm-linux-gnueabihf-gcc")
set(CMAKE_CXX_COMPILER "arm-linux-gnueabihf-g++")
else()
message(FATAL_ERROR "Unknown arch abi ${TARGET_ARCH_ABI}, only support armv8 and armv7hf.")
return()
endif()
project(yolov3_detection)
message(STATUS "TARGET ARCH ABI: ${TARGET_ARCH_ABI}")
message(STATUS "PADDLE LITE DIR: ${PADDLE_LITE_DIR}")
include_directories(${PADDLE_LITE_DIR}/include)
link_directories(${PADDLE_LITE_DIR}/libs/${TARGET_ARCH_ABI})
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
if(TARGET_ARCH_ABI STREQUAL "armv8")
set(CMAKE_CXX_FLAGS "-march=armv8-a ${CMAKE_CXX_FLAGS}")
set(CMAKE_C_FLAGS "-march=armv8-a ${CMAKE_C_FLAGS}")
elseif(TARGET_ARCH_ABI STREQUAL "armv7hf")
set(CMAKE_CXX_FLAGS "-march=armv7-a -mfloat-abi=hard -mfpu=neon-vfpv4 ${CMAKE_CXX_FLAGS}")
set(CMAKE_C_FLAGS "-march=armv7-a -mfloat-abi=hard -mfpu=neon-vfpv4 ${CMAKE_C_FLAGS}" )
endif()
find_package(OpenMP REQUIRED)
if(OpenMP_FOUND OR OpenMP_CXX_FOUND)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OpenMP_C_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OpenMP_CXX_FLAGS}")
message(STATUS "Found OpenMP ${OpenMP_VERSION} ${OpenMP_CXX_VERSION}")
message(STATUS "OpenMP C flags: ${OpenMP_C_FLAGS}")
message(STATUS "OpenMP CXX flags: ${OpenMP_CXX_FLAGS}")
message(STATUS "OpenMP OpenMP_CXX_LIB_NAMES: ${OpenMP_CXX_LIB_NAMES}")
message(STATUS "OpenMP OpenMP_CXX_LIBRARIES: ${OpenMP_CXX_LIBRARIES}")
else()
message(FATAL_ERROR "Could not found OpenMP!")
return()
endif()
find_package(OpenCV REQUIRED)
if(OpenCV_FOUND OR OpenCV_CXX_FOUND)
include_directories(${OpenCV_INCLUDE_DIRS})
message(STATUS "OpenCV library status:")
message(STATUS " version: ${OpenCV_VERSION}")
message(STATUS " libraries: ${OpenCV_LIBS}")
message(STATUS " include path: ${OpenCV_INCLUDE_DIRS}")
else()
message(FATAL_ERROR "Could not found OpenCV!")
return()
endif()
add_executable(yolov3_detection yolov3_detection.cc)
target_link_libraries(yolov3_detection paddle_light_api_shared ${OpenCV_LIBS})
target_link_libraries(yolov3_detection wiringPi)#!/bin/bash
# configure
TARGET_ARCH_ABI=armv8 # for RK3399, set to default arch abi
#TARGET_ARCH_ABI=armv7hf # for Raspberry Pi 3B
PADDLE_LITE_DIR=/home/pi/paddle-lite/build.lite.armlinux.armv8.gcc/inference_lite_lib.armlinux.armv8/cxx
if [ "x$1" != "x" ]; then
TARGET_ARCH_ABI=$1
fi
# build
rm -rf build
mkdir build
cd build
cmake -DPADDLE_LITE_DIR=${PADDLE_LITE_DIR} -DTARGET_ARCH_ABI=${TARGET_ARCH_ABI} ..
make
# LD_LIBRARY_PATH=$LD_LIBRARY_PATH:${PADDLE_LITE_DIR}/libs/ ./yolov3_detection ppyolo18.nb /home/pi/PaddleX/deploy/raspberry/hard_hat_workers0.jpgpi@raspberrypi:~/paddle-lite/lite/demo/cxx/yolov3_detection $ sh build.sh
-- The C compiler identification is GNU 8.3.0
-- The CXX compiler identification is GNU 8.3.0
-- Check for working C compiler: /usr/bin/aarch64-linux-gnu-gcc
-- Check for working C compiler: /usr/bin/aarch64-linux-gnu-gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/aarch64-linux-gnu-g++
-- Check for working CXX compiler: /usr/bin/aarch64-linux-gnu-g++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- TARGET ARCH ABI: armv8
-- PADDLE LITE DIR: /home/pi/paddle-lite/build.lite.armlinux.armv8.gcc/inference_lite_lib.armlinux.armv8/cxx
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- Found OpenMP 4.5
-- OpenMP C flags: -fopenmp
-- OpenMP CXX flags: -fopenmp
-- OpenMP OpenMP_CXX_LIB_NAMES: gomp;pthread
-- OpenMP OpenMP_CXX_LIBRARIES: /usr/lib/gcc/aarch64-linux-gnu/8/libgomp.so;/usr/lib/aarch64-linux-gnu/libpthread.so
-- Found OpenCV: /usr (found version "3.2.0")
-- OpenCV library status:
-- version: 3.2.0
-- libraries: opencv_calib3d;opencv_core;opencv_features2d;opencv_flann;opencv_highgui;opencv_imgcodecs;opencv_imgproc;opencv_ml;opencv_objdetect;opencv_photo;opencv_shape;opencv_stitching;opencv_superres;opencv_video;opencv_videoio;opencv_videostab;opencv_viz;opencv_aruco;opencv_bgsegm;opencv_bioinspired;opencv_ccalib;opencv_datasets;opencv_dpm;opencv_face;opencv_freetype;opencv_fuzzy;opencv_hdf;opencv_line_descriptor;opencv_optflow;opencv_phase_unwrapping;opencv_plot;opencv_reg;opencv_rgbd;opencv_saliency;opencv_stereo;opencv_structured_light;opencv_surface_matching;opencv_text;opencv_ximgproc;opencv_xobjdetect;opencv_xphoto
-- include path: /usr/include;/usr/include/opencv
-- Configuring done
-- Generating done
-- Build files have been written to: /home/pi/paddle-lite/lite/demo/cxx/yolov3_detection/build
Scanning dependencies of target yolov3_detection
[ 50%] Building CXX object CMakeFiles/yolov3_detection.dir/yolov3_detection.cc.o
[100%] Linking CXX executable yolov3_detection
/usr/bin/ld: 当搜索用于 /usr/lib/gcc/aarch64-linux-gnu/8/../../../../lib/libwiringPi.so 时跳过不兼容的 -lwiringPi
/usr/bin/ld: 当搜索用于 /usr/lib/../lib/libwiringPi.so 时跳过不兼容的 -lwiringPi
/usr/bin/ld: 当搜索用于 /usr/lib/gcc/aarch64-linux-gnu/8/../../../libwiringPi.so 时跳过不兼容的 -lwiringPi
[100%] Built target yolov3_detectionpi@raspberrypi:~/paddle-lite/lite/demo/cxx/yolov3_detection $ ./build/yolov3_detection det_yolov3_mobilenetv3.nb
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1064 Setup] ARM multiprocessors name: HARDWARE : BCM2835
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1065 Setup] ARM multiprocessors number: 4
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1067 Setup] ARM multiprocessors ID: 0, max freq: 1200, min freq: 1200, cluster ID: 0, CPU ARCH: A53
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1067 Setup] ARM multiprocessors ID: 1, max freq: 1200, min freq: 1200, cluster ID: 0, CPU ARCH: A53
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1067 Setup] ARM multiprocessors ID: 2, max freq: 1200, min freq: 1200, cluster ID: 0, CPU ARCH: A53
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1067 Setup] ARM multiprocessors ID: 3, max freq: 1200, min freq: 1200, cluster ID: 0, CPU ARCH: A53
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1073 Setup] L1 DataCache size is:
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1075 Setup] 32 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1075 Setup] 32 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1075 Setup] 32 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1075 Setup] 32 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1077 Setup] L2 Cache size is:
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1079 Setup] 512 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1079 Setup] 512 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1079 Setup] 512 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1079 Setup] 512 KB
[I 12/ 1 23:46:20.593 .../pi/paddle-lite/lite/core/device_info.cc:1081 Setup] L3 Cache size is:
[I 12/ 1 23:46:20.594 .../pi/paddle-lite/lite/core/device_info.cc:1083 Setup] 0 KB
[I 12/ 1 23:46:20.594 .../pi/paddle-lite/lite/core/device_info.cc:1083 Setup] 0 KB
[I 12/ 1 23:46:20.594 .../pi/paddle-lite/lite/core/device_info.cc:1083 Setup] 0 KB
[I 12/ 1 23:46:20.594 .../pi/paddle-lite/lite/core/device_info.cc:1083 Setup] 0 KB
[I 12/ 1 23:46:20.594 .../pi/paddle-lite/lite/core/device_info.cc:1085 Setup] Total memory: 881860KB
detection, image size: 640, 480, detect object: head, score: 0.917295, location: x=155, y=25, width=128, height=164
Preprocess time: 8424.817383 ms
detection, image size: 640, 480, detect object: head, score: 0.899408, location: x=157, y=23, width=125, height=166
Preprocess time: 8228.533203 ms
detection, image size: 640, 480, detect object: head, score: 0.902913, location: x=156, y=25, width=126, height=161
Preprocess time: 8470.201172 ms
仅就mAP这个评估指标而言，yolov3_mobilenetv3远胜于ssd_mobilenet_v1: 0.85+ v.s. 0.4+
但预测时间这个指标就比较惨了，yolov3_mobilenetv3远远落后于ssd_mobilenet_v1: 8000+ ms v.s. 300 ms
后续将继续关注ppyolo等其它模型的部署效果，敬请期待。