StdTracker_Board_App/sophon_bm1684x/report/DualTest.cpp

#include <iostream>
#include <cstdlib>
#include <string>
#include <fstream>
#include <dirent.h>
#include <unistd.h>
#include <sys/stat.h>
#include <thread>
#include "ff_decode.hpp"
#include <cstdio>
#include <cstdint>
#include "demo_utils.h"
#include "Arith_Nano_Tracker_API.h"
#include "detection_type_api.h"
#include "NeoArithStandardDll.h"
#include "AI_API.h"
#include "threadpool.h"

static int DETECT_MODEL_TYPE = 0;
static int IMAGEWIDTH        = 0;
static int IMAGEHEIGHT       = 0;

TenMat * maskImg = (TenMat*)malloc(sizeof(TenMat));
void * maskdata = (void*)malloc(IMAGEWIDTH*IMAGEHEIGHT*sizeof(int64_t));


std::vector<objinfo> visdets1;
std::vector<objinfo> visdets2;

void AI_Det_Release()
{
    int ret = Uninit_AI();
    if (maskdata!=NULL)
        free(maskdata);
    if (maskImg!=NULL)
        free(maskImg);
    printf ("Uninit_AI DETECTER Sucess!\n");
}

void AI_Det_Init(int model_type)
{
    if (model_type == 0)
    {
        IMAGEWIDTH = 1280;
        IMAGEHEIGHT = 1024;
    }
    else
    {
        IMAGEWIDTH = 640;
        IMAGEHEIGHT = 512;
    }
    
    std::string visdetmodeFull = "../models/Detect/1684x/vis_common_headers_1280x1024_bm1684x_F16_false.bmodel";
    std::string visdetmodePart = "../models/Detect/1684x/vis_common_headers_640x512_bm1684x_F16_false.bmodel";

	int ret = 0;
    ret = Init_ACL();
    if(ret){
        printf("Init_ACL error\n");
    }
    ret = set_deque_size(5,5,5,5,5,5,5,5,9,9,9,9);
    if(ret){
        printf("set_deque_size error\n");
    }
    printf("deque setted\n");
    // 设置红外图像分辨率
    ret = set_stream_size(IMAGEWIDTH, IMAGEHEIGHT, IR);
    if(ret){
        printf("set_stream_size error\n");
    }
    // 设置可见光图像分辨率
    ret = set_stream_size(IMAGEWIDTH, IMAGEHEIGHT, VIS);
    if(ret){
        printf("set_stream_size error\n");
    }
    //设置类别标签， 标签标请咨询对应的算法工程师
    uint32_t modelclass[] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14};
    uint32_t standerclass[] = {10000, 20000, 30000, 40000, 50000,
                               60000, 70000, 80000, 90000, 100000,
                               110000, 120000, 130000, 140000, 150000};
    ret = setLabelMap(modelclass, standerclass, sizeof(modelclass)/sizeof(uint32_t));

    if(ret){
        printf("setLabelMap error with ret=%d\n", ret);
    }
    // 设置数据流格式，NV12 or RGB
    ret = set_StreamMode(AIGO::INPUT_LAYOUT::NV12);
    if(ret){
        printf("set_StreamMode error with ret=%d\n", ret);
    }
    // 初始化模型
    if (model_type == 0){
        ret = Init_AI(visdetmodeFull.c_str(), visdetmodeFull.c_str(), NULL, NULL);

    }else{
        ret = Init_AI(visdetmodeFull.c_str(), visdetmodePart.c_str(), NULL, NULL);
    }
    if(ret){
        printf("init error\n");
    }

    ret = set_dropout_param(false);
    if(ret){
        printf("set dropout error\n");
    }
    // ret = switch_vis_to_ir();
    printf("get det threshold\n");
    int cNum = sizeof(modelclass)/sizeof(uint32_t);
    for (int z = 0; z < cNum; ++z) {
        ret = set_det_threshold(0.45, standerclass[z]);
    }

    if(ret){
        printf("set det threshold error\n");
    }

    set_dropout_param(1);

}



void AI_Det_Run(cv::Mat NV12Mat,int frnID)
{
    //printf("start ai run\n");
    ImgMat imgvis;
    imgvis.data = NV12Mat.data;
    imgvis.width = IMAGEWIDTH;
    imgvis.height = IMAGEHEIGHT;
    imgvis.depth = 1;
    imgvis.layout = HWC;
    imgvis.inlayout = AIGO::INPUT_LAYOUT::NV12;
    imgvis.mode = VIS;
    imgvis.timestamp = frnID;
    long long timestamp = frnID;

    int ret = runAsync(imgvis, true, false, false);
    if(ret)
    {
        printf("runAsync error\n");
    }

    // 按照红外再送一次推理
    ImgMat imgir = imgvis;
    imgir.mode = IR;
    
    ret = runAsync(imgir, true, false, false);
    if(ret)
    {
        printf("runAsync error\n");
    }

    maskImg->data = maskdata;
    visdets1.clear();
    ret = get_result(timestamp, visdets1, maskImg, false, IR);

    visdets2.clear();
    ret = get_result(timestamp, visdets2, maskImg, false, VIS);

    printf("end ai run\n");

    printf ("**** F:[%d] AIOD_Detect num:[%d] [%d]  timestamp: %d\n",frnID,visdets1.size(),visdets2.size(), timestamp);
}


void draw(cv::Mat& frame,ARIDLL_OUTPUT output)
{
    auto trackerOut = output.stTrackers[0];
    auto detOut_array = output.stAlarmObjs;
    auto detOut_cnt = output.nAlarmObjCnts;

    for (size_t i = 0; i < detOut_cnt; i++)
    {
        cv::Rect outRect;
        outRect.width = MAX(15,(int)detOut_array[i].nObjW);
        outRect.height= MAX(15, (int)detOut_array[i].nObjH);
        outRect.x = (int)detOut_array[i].nX-outRect.width/2;
        outRect.y = (int)detOut_array[i].nY-outRect.height/2;
        cv::rectangle(frame,outRect,cv::Scalar(255,0,0),3);
        cv::putText(frame,to_string(detOut_array[i].nOutputID),cv::Point(outRect.x - 10,outRect.y),
        cv::FONT_HERSHEY_COMPLEX, 0.5, cv::Scalar(255,0,0));
    }
    
    cv::Rect outRect;
    outRect.width = MAX(15,(int)trackerOut.nObjW);
    outRect.height= MAX(15, (int)trackerOut.nObjH);
    outRect.x = (int)trackerOut.nX-outRect.width/2;
    outRect.y = (int)trackerOut.nY-outRect.height/2;
    cv::rectangle(frame,outRect,cv::Scalar(0,0,255),5);
}

// 读取1个视频按照两路执行，测试双路识别跟踪
int main(int argc, char** argv)
{
    threadpool executor{ 4 };

    cv::VideoCapture capture(argv[2]);
	int nWidth = int(capture.get(cv::CAP_PROP_FRAME_WIDTH));
	int nHeight = int(capture.get(cv::CAP_PROP_FRAME_HEIGHT));

	cv::VideoWriter writer;
	Size size = cv::Size(nWidth * 2, nHeight);
    bool bflag = writer.open("trackbm1684x.avi", cv::VideoWriter::fourcc('M', 'J', 'P', 'G'), 25, size, true);

	printf("nW = %d,nH = %d \n",nWidth,nHeight);

    // 初始化识别:内部处理双路
    DETECT_MODEL_TYPE = *argv[1];
    AI_Det_Init(DETECT_MODEL_TYPE);

    printf("ai detect init done!");

    // 创建跟踪算法句柄
	ArithHandle pTracker1 = STD_CreatEOArithHandle();
    ArithHandle pTracker2 = STD_CreatEOArithHandle();

    // 初始化为凝视-对地模式
	ARIDLL_EOArithInitWithMode(pTracker1,nWidth,nHeight,GD_PIXEL_FORMAT_E::GD_PIXEL_FORMAT_GRAY_Y16,
	GLB_SYS_MODE::GLB_SYS_STARE,GLB_SCEN_MODE::GLB_SCEN_GROUND);
    ARIDLL_EOArithInitWithMode(pTracker2,nWidth,nHeight,GD_PIXEL_FORMAT_E::GD_PIXEL_FORMAT_GRAY_Y16,
	GLB_SYS_MODE::GLB_SYS_STARE,GLB_SCEN_MODE::GLB_SCEN_GROUND);

    // 算法输入部分
	ARIDLL_INPUTPARA stInputPara = { 0 };
	stInputPara.unFrmId++;
	stInputPara.stCameraInfo.fPixelSize = 15;
	stInputPara.stCameraInfo.nFocus = 300;

    // 算法输出部分
	ARIDLL_OUTPUT stOutput1 = { 0 };
    ARIDLL_OUTPUT stOutput2 = { 0 };

	unsigned char* pY8Data = new unsigned char[nWidth * nHeight];


    FILE* file = fopen("timer.log","w+");


    // 模拟算法执行流程
	int nTrackSuc = 0;
	cv::Mat frame;
	while(1)
	{
		stInputPara.unFrmId++;

		capture.read(frame);
        if(frame.empty())
            break;
        int imgW = frame.cols;  // img_w
        int imgH = frame.rows; // img_h


		cv::Mat srcGray(nHeight,nWidth,CV_8UC1);
		cv::cvtColor(frame,srcGray,cv::COLOR_RGB2GRAY);
		for (size_t i = 0; i < nWidth * nHeight; i++)
		{
			pY8Data[i] = srcGray.data[i];
		}

		cv::Mat resizedImg;
        cv::resize(frame, resizedImg, cv::Size(IMAGEWIDTH, IMAGEHEIGHT));
        cv::Mat NV12Mat;
        BGR2YUV_nv12(resizedImg, NV12Mat);

        cv::TickMeter tm;
		tm.start();

        // 调用AI识别异步接口
        AI_Det_Run(NV12Mat,stInputPara.unFrmId);

        // remap 到原图
        for (size_t i = 0; i < visdets1.size(); i++)
        {
            auto ratioW = (imgW *1.0) / IMAGEWIDTH;
            auto ratioH = (imgH *1.0) / IMAGEHEIGHT;
            visdets1[i].x1 *= ratioW;
            visdets1[i].x2 *= ratioW;
            visdets1[i].y1 *= ratioH;
            visdets1[i].y2 *= ratioH;
        }
        for (size_t i = 0; i < visdets2.size(); i++)
        {
            auto ratioW = (imgW *1.0) / IMAGEWIDTH;
            auto ratioH = (imgH *1.0) / IMAGEHEIGHT;
            visdets2[i].x1 *= ratioW;
            visdets2[i].x2 *= ratioW;
            visdets2[i].y1 *= ratioH;
            visdets2[i].y2 *= ratioH;
        }

        ARIDLL_MergeAITargets(pTracker1,visdets1.data(),visdets1.size());
        ARIDLL_MergeAITargets(pTracker2,visdets2.data(),visdets2.size());


        // 下发面锁定指令
		if (stInputPara.unFrmId == 10)
		{
			ARIDLL_LockCommand(pTracker1, 228*2,279*2,40,25);
		}

		if (stInputPara.unFrmId == 300)
		{
			ARIDLL_LockCommand(pTracker2, 228*2,279*2,40,25);
		}


        // 构建图像类型
		GD_VIDEO_FRAME_S img = { 0 };
		img.enPixelFormat = GD_PIXEL_FORMAT_E::GD_PIXEL_FORMAT_GRAY_Y8;
		img.u32Width = nWidth;
		img.u32Height = nHeight;
		img.u32Stride[0] = img.u32Width * 1;
		img.u64VirAddr[0] = (unsigned char*)pY8Data;



        // 运行算法主控逻辑API
        std::future<int> out1 = executor.commit("Tracker1",ARIDLL_RunController,pTracker1, img, stInputPara, &stOutput1);
		std::future<int> out2 = executor.commit("Tracker2",ARIDLL_RunController,pTracker2, img, stInputPara, &stOutput2);
        out1.get();
        out2.get();

		tm.stop();

		printf("time:%.2f\n",tm.getTimeMilli());


        fprintf(file,"%.1f\n",tm.getTimeMilli());


        draw(frame,stOutput1);
        auto frame2 = frame.clone();
        draw(frame2,stOutput2);

        cv::Mat contact;
	    cv::hconcat(frame, frame2, contact);


        cv::putText(contact,"bm1684x fps:" + to_string(1000/tm.getTimeMilli()),cv::Point(100,50),
        cv::FONT_HERSHEY_COMPLEX, 6, cv::Scalar(255,0,0));


		//writer << contact;
	}

    

    AI_Det_Release();

	delete[] pY8Data;

	writer.release();
}