两百行C++代码实现yolov5人形运动检测

简单实现了一下yolov5单个人形的运动检测功能，多个的话涉及到多目标跟踪，以后有空再弄。实现思路即目标检测+静止抑制，具体步骤如下：
（1）训练yolov5模型，这里就没有自己训练了，直接使用官方的开源模型yolov5s.pt；
（2）运行yolov5工程下面的export.py，将pt模型转成onnx模型；
（3）编写yolov5部署的C++工程，包括前处理、推理和后处理部分；
（4）设置一个固定长度的队列，用来存放每帧检测出单个人形检测框的中心点；
（5）读取视频第一帧，用yolov5检测第一帧图像的人形，检测器能可能检测出0或多个人形，如果检测数不为1则直接跳过读取后续帧；
（6）读取后续帧，直到检测出人形的帧数达到队列长度，此时队列已满；
（7）此时再处理图片帧的时候，需要弹出队列首，把新计算得到的检测框中心点添加到队列尾；
（8）计算队列中相邻元素检测框中心点的距离，若相邻元素检测框中心点的距离都大于阈值，说明检测物在一直移动，显示运动检测框，否则不显示。
具体实现的时候，队列的长度和移动距离的阈值需要调节，LZ取队列长度为5，移动距离阈值为检测框对角线长度除以100。下面的代码基本上实现了实时运动检测功能，但不知道有什么应用哈哈，LZ就疫情在家没事写了玩的。

#include <iostream>
#include <fstream>
#include <opencv2/opencv.hpp>


#define DEBUG


// 常量
const float INPUT_WIDTH = 640.0;
const float INPUT_HEIGHT = 640.0;
const float SCORE_THRESHOLD = 0.6;
const float NMS_THRESHOLD = 0.25;
const float CONFIDENCE_THRESHOLD = 0.6;

const std::vector<std::string> class_name = {
	"person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck", "boat", "traffic light",
	"fire hydrant", "stop sign", "parking meter", "bench", "bird", "cat", "dog", "horse", "sheep", "cow",
	"elephant", "bear", "zebra", "giraffe", "backpack", "umbrella", "handbag", "tie", "suitcase", "frisbee",
	"skis", "snowboard", "sports ball", "kite", "baseball bat", "baseball glove", "skateboard", "surfboard",
	"tennis racket", "bottle", "wine glass", "cup", "fork", "knife", "spoon", "bowl", "banana", "apple",
	"sandwich", "orange", "broccoli", "carrot", "hot dog", "pizza", "donut", "cake", "chair", "couch",
	"potted plant", "bed", "dining table", "toilet", "tv", "laptop", "mouse", "remote", "keyboard", "cell phone",
	"microwave", "oven", "toaster", "sink", "refrigerator", "book", "clock", "vase", "scissors", "teddy bear",
	"hair drier", "toothbrush" };

const int QUEUE_SIZE = 5;


//画出检测框和标签
void draw_label(cv::Mat& input_image, std::string label, int left, int top)
{
	int baseLine;
	cv::Size label_size = cv::getTextSize(label, 0.7, 0.7, 1, &baseLine);
	top = std::max(top, label_size.height);
	cv::Point tlc = cv::Point(left, top);
	cv::Point brc = cv::Point(left, top + label_size.height + baseLine);
	cv::putText(input_image, label, cv::Point(left, top + label_size.height), cv::FONT_HERSHEY_SIMPLEX, 0.7, cv::Scalar(0, 255, 255), 1);
}


//预处理
std::vector<cv::Mat> preprocess(cv::Mat & input_image, cv::dnn::Net & net)
{
	cv::Mat blob;
	cv::dnn::blobFromImage(input_image, blob, 1. / 255., cv::Size(INPUT_WIDTH, INPUT_HEIGHT), cv::Scalar(), true, false);

	net.setInput(blob);

	std::vector<cv::Mat> preprcess_image;
	net.forward(preprcess_image, net.getUnconnectedOutLayersNames());

	return preprcess_image;
}


//后处理
std::vector<cv::Rect> postprocess(std::vector<cv::Mat> & preprcess_image, cv::Mat & output_image, int class_index)
{
	std::vector<int> class_ids;
	std::vector<float> confidences;
	std::vector<cv::Rect> boxes;
	std::vector<cv::Rect> boxes_nms;

	float x_factor = output_image.cols / INPUT_WIDTH;
	float y_factor = output_image.rows / INPUT_HEIGHT;

	float* data = (float*)preprcess_image[0].data;

	const int dimensions = 85;
	const int rows = 25200;
	for (int i = 0; i < rows; ++i)
	{
		float confidence = data[4];
		if (confidence >= CONFIDENCE_THRESHOLD)
		{
			float* classes_scores = data + 5;
			cv::Mat scores(1, class_name.size(), CV_32FC1, classes_scores);
			cv::Point class_id;
			double max_class_score;
			cv::minMaxLoc(scores, 0, &max_class_score, 0, &class_id);
			if (max_class_score > SCORE_THRESHOLD)
			{
				confidences.push_back(confidence);
				class_ids.push_back(class_id.x);

				float cx = data[0];
				float cy = data[1];
				float w = data[2];
				float h = data[3];
				int left = int((cx - 0.5 * w) * x_factor);
				int top = int((cy - 0.5 * h) * y_factor);
				int width = int(w * x_factor);
				int height = int(h * y_factor);
				boxes.push_back(cv::Rect(left, top, width, height));
			}
		}
		data += 85;
	}

	std::vector<int> indices;
	cv::dnn::NMSBoxes(boxes, confidences, SCORE_THRESHOLD, NMS_THRESHOLD, indices);
	for (size_t i = 0; i < indices.size(); i++)
	{
		int idx = indices[i];
		if (class_ids[idx] == class_index)
		{
			cv::Rect box = boxes[idx];
			boxes_nms.push_back(box);

			int left = box.x;
			int top = box.y;
			int width = box.width;
			int height = box.height;

			std::string label = cv::format("%.2f", confidences[idx]);
			label = class_name[class_ids[idx]] + ":" + label;
			draw_label(output_image, label, left, top);
		}
	}

	return boxes_nms;
}


//计算检测框的中心
cv::Point get_center(cv::Rect detection)
{
	return cv::Point(detection.x + detection.width / 2, detection.y + detection.height / 2);;
}


//计算两点间距离
float get_distance(cv::Point p1, cv::Point p2)
{
	return sqrt(pow(p1.x - p2.x, 2) + pow(p1.y - p2.y, 2));
}



int main(int argc, char** argv)
{
	cv::VideoCapture capture(0);
	cv::Mat frame;
	cv::dnn::Net net = cv::dnn::readNet("yolov5s-f32.onnx");

	int frame_num = 0;
	std::vector<cv::Point> detection_center_queue;

	while (cv::waitKey(1) < 0)
	{
		capture >> frame;
		if (frame.empty())
			break;

		std::cout << "******************************************************************* frame_num: " << frame_num << std::endl;

		cv::Mat image = frame.clone();
		std::vector<cv::Mat> preprcess_image = preprocess(image, net);

		std::vector<cv::Rect> detections = postprocess(preprcess_image, image, 0);
		cv::Rect detection;

		if (detections.size() != 1)
			goto FLAG;

		detection = detections[0];

		if (frame_num < QUEUE_SIZE)
		{
			cv::Point detection_center = get_center(detection);
			detection_center_queue.push_back(detection_center);

#ifdef DEBUG
			std::cout << "detection_center_queue:" << std::endl;
			for (size_t i = 0; i < detection_center_queue.size(); i++)
			{
				std::cout << detection_center_queue[i] << " ";
			}
			std::cout << std::endl;
#endif // DEBUG
		}
		else
		{
			detection_center_queue.erase(detection_center_queue.begin());
			cv::Point detection_center = get_center(detection);
			detection_center_queue.push_back(detection_center);
			
#ifdef DEBUG
			std::cout << "detection_center_queue:" << std::endl;
			for (size_t i = 0; i < detection_center_queue.size(); i++)
			{
				std::cout << detection_center_queue[i] << " ";
			}
			std::cout << std::endl;
#endif // DEBUG

			std::vector<float> distances;

			for (size_t i = 1; i < detection_center_queue.size(); i++)
			{
				cv::Point detection_center_old = detection_center_queue[i - 1];
				cv::Point detection_center_new = detection_center_queue[i];

				float distance = get_distance(detection_center_new, detection_center_old);
				distances.push_back(distance);
				std::cout << distance << std::endl;
			}

			bool show_box = true;
			for (size_t i = 0; i < distances.size(); i++)
			{
				float moving_threshold = sqrt(pow(detection.width, 2) + pow(detection.height, 2)) / 100;
				//float moving_threshold = 5;
				if (distances[i] < moving_threshold)
					show_box = false;
			}

			if(show_box)
				cv::rectangle(image, cv::Point(detection.x, detection.y), \
					cv::Point(detection.x + detection.width, detection.y + detection.height), cv::Scalar(0, 0, 255), 2);
		}

		frame_num++;

	FLAG:
		cv::imshow("output", image);		
	}

	capture.release();
	return 0;
}

后续：多目标跟踪版本的实现sort+yolov5