Tuo Wenbo / WeiQ_Mianfang

using OpenCvSharp;
using OS.Spin.BusinessLayer.SubBusiness;
using OS.Spin.Common;
using OS.Spin.Common.Files;
using OS.Spin.Common.MultiThread;
using OS.Spin.Modle.BusinessLayer;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Runtime.InteropServices;
using System.Threading;
using static OS.Spin.Modle.Sdk.StructInfos;

namespace OS.Spin.BusinessLayer.MainBusiness
{
    public sealed class FlawDetectionBl
    {
        #region 变量
        private int _cameraCount = 0;
        private List<CameraBl> _cameras = new List<CameraBl>();
        private List<DetectionBl> _detections = new List<DetectionBl>();
        private Thread _runThread;
        private SemaphorePool _semaphorePool;
        private string _classId = string.Empty;
        private int _imageId = 0;
        private int _reallyImageId = 0;
        private int _finishedCameraCount = 0;
        private IntPtr _initHandle = new IntPtr(0);
        private IntPtr _handle = new IntPtr(1);
        private const int RS_FACTOR = 1;
        private const int ROI_WIDTH = 500;
        private const int ROI_HEIGHT = 250;
        public delegate void WindowsControl(int cmd);
        public WindowsControl TlWorker;
        private IntPtr _handle_5 = new IntPtr(5);

        private IntPtr _handle_37 = new IntPtr(37);

        private IntPtr _handle_h = new IntPtr(12);
        private IntPtr _handle_v = new IntPtr(22);
        private IntPtr _handle_proc = new IntPtr(32);

        private Modle.Sdk.StructInfos.FinalResults[] label_poolarray = new Modle.Sdk.StructInfos.FinalResults[100];
        private Modle.Sdk.StructInfos.FinalResults[] final_labelout = new Modle.Sdk.StructInfos.FinalResults[16];
        //private int pool_size = 0;

        private int _gainValue = 0;

        private bool _isSetGain = false;

        private int current_meter = 0;
        //private MachineStopBl _stopBl = null;
        //private MaxSmallTransferBl _transferBl = null;
        private bool _canRecived = true;
        //private string[] _lableNames = new[] { "背景", "停车痕(紧)", "停车痕(松)", "断经", "错花", "并纬", "缩纬", "缺纬", "糙纬", "折返", "断纬",
        //    "油污", "浆斑", "污经","经条", "擦白", "擦伤","空织","破洞","紧经","污纬","皱印","上首破边","下首破边", "刮伤", "纬密变化","蛛网","挡车工标记","毛纱、线头","起皱","水印","水波纹", "其他" };
        private string[] _lableNames = new[] { "背景","棉花","大棉团","横向折痕","经向折痕","小竹节","杂物","棉球","结头","破洞","拖纱","断疵","跳花","云织","经竹节","经缩","粗经","经向异纤","起圈纬缩","纬竹节","纬缩",
            "松纬缩","纬向异纤","断经_黑","断经_亮","松经","稀纬","稀弄","双纬","空织","百脚","密路","花毛密路","错支类","豁边","紧边","烂边毛边","深色油污","透明油污","交班印","条干不匀","其他瑕疵","人为干扰" };

        public delegate void AddFlawShot(MFlawInfo flaw);
        public AddFlawShot DoAddFlawShot;

        public delegate void RunningEvent();
        public RunningEvent Running;
        private int _lostTime = 0;
        private TimerBl _tbl = null;

        private Modle.Sdk.StructInfos.CLAS_PARAM _findWall;

        #endregion

        #region 程序入口
        public FlawDetectionBl(int cameraCount)
        {
            try
            {
                _cameraCount = cameraCount;
                _finishedCameraCount = cameraCount;
                LogisTrac.WriteInfoLog(String.Format("[开始] =====[{0}]======", "FlawDetectionBl"));
                _findWall = new Modle.Sdk.StructInfos.CLAS_PARAM
                {
                    prob_cls = ConfigHelper.GetInstance().GetConfig().Prob_cls,//通用瑕疵
                    prob_sma = ConfigHelper.GetInstance().GetConfig().Prob_sma,//小瑕疵
                    prob_color = ConfigHelper.GetInstance().GetConfig().Prob_color,//色纤
                    prob_thin = ConfigHelper.GetInstance().GetConfig().Prob_thin,//淡痕
                    len_sma = ConfigHelper.GetInstance().GetConfig().Len_sma,//小瑕疵长度限制阈值
                    cloths_kind = 0,   //布种 0:正常布种; 1:薄布 2:厚布
                    m_gain = 0,     //增益返回值
                    m_base = ConfigHelper.GetInstance().GetConfig().M_base, //增益基准值
                    m_meters = (int)OS.Spin.Running.Cache.GetInstance().CMeter,    //当前米数计数值
                    thre_are = ConfigHelper.GetInstance().GetConfig().Thre_Area,//点状瑕疵面积
                    thre_location = ConfigHelper.GetInstance().GetConfig().Thre_Location//布尾米数位置阈值
                };

                #region TWB20210902
                //// 初始化分类
                var count = SdkLayer.SdkMyCode.Flaw_ClassifyInit(ref _handle_h, ref _handle_v, ref _handle_proc);
                #endregion
                // 初始化后处理
                //count = OS.Spin.SdkLayer.SdkImport.Detc_Post_Init(ref _handle, _cameraCount);
                _classId = Guid.NewGuid().ToString();
                // 获取信号量
                _semaphorePool = OS.Spin.Common.MultiThread.SemaphorePool.GetInstance();
                // 创建检查信号量
                _semaphorePool.MakeSemaphore(string.Format("{0}{1}", OS.Spin.Running.Infos.Names.S_DETECTION_OK, _classId));

                // 初始化相机
                for (var cId = 0; cId < _cameraCount; cId++)
                {
                    var camera = new CameraBl(cId)
                    {
                        SendRecivedImg = RecivedImg
                    };

                    _cameras.Add(camera);

                    DetectionBl detection = new DetectionBl(_classId, cId)
                    {
                        DoGetThreeChannelMat = DoGetThreeChannelsMat
                    };
                    _detections.Add(detection);
                }

                // 启动分类线程
                if (null == _runThread)
                {
                    _runThread = new Thread(Classifying);
                    _runThread.Priority = ThreadPriority.AboveNormal;
                    _runThread.Start();
                }

                //// 创建码表通讯模块
                //_tbl = new TimerBl
                //{
                //    DoRecived = TimerRecived
                //};
                //_tbl.OnStart();
            }
            catch (Exception ex)
            {
                LogisTrac.WriteLog(string.Format("FlawDetectionBl:{0}", ex.Message));
            }

        }

        public bool CanTrigger
        {
            get { return _canRecived; }
        }

        private FlawDetectionBl()
        {

        }
        #endregion

        #region public

        #region 相机触发拍照
        /// <summary>
        /// 相机触发拍照
        /// </summary>
        /// <returns></returns>
        public bool CameraTrigger(double meter)
        {
            try
            {
                LogisTrac.WriteLog("触发");
                if (meter < 3)
                {
                    foreach (var camera in _cameras)
                    {
                        camera.SetCameraCainValue(ConfigHelper.GetInstance().GetConfig().M_base / 10.0);
                        LogisTrac.WriteLog(string.Format("5米前设置相机增益===[{0}]=====", ConfigHelper.GetInstance().GetConfig().M_base / 10.0));
                    }
                    _isSetGain = true;
                    //pool_size = 0;
                }
                OS.Spin.Running.Cache.GetInstance().CMeter = meter;
                Running?.Invoke();
                if (!_canRecived)
                {
                    _lostTime++;
                    if (_lostTime > 10)
                    {
                        _canRecived = true;
                    }
                    return false;
                }
                _lostTime = 0;
                _canRecived = false;
                OS.Spin.Running.Cache.GetInstance().CMeter = meter;
                //if (_finishedCameraCount >= _cameraCount)
                {
                    //_finishedCameraCount = 0;
                    // 重置相机接收完成标志
                    //OS.Spin.Running.Cache.GetInstance().CamerasOk = false;
                    if (meter > 3 && _isSetGain)
                    {
                        foreach (var camera in _cameras)
                        {
                            camera.SetCameraCainValue((_gainValue / 10.0));
                        }
                        _isSetGain = false;
                        //_gainValue = 0;
                        LogisTrac.WriteLog(string.Format("设置相机增益值===[{0}]==============", _gainValue));
                    }

                    foreach (var camera in _cameras)
                    {
                        camera.Trigger();
                    }
                    return true;
                }
            }
            catch (Exception ex)
            {
                LogisTrac.WriteLog(string.Format("CameraTrigger:{0}", ex.Message));
                return false;
            }

            //return false;
        }
        #endregion

        #region 重置清零后处理参数
        /// <summary>
        /// 重置清零后处理参数
        /// </summary>
        public void ReSetDp()
        {
            //OS.Spin.SdkLayer.SdkImport.DP_Clear(_handle);
        }
        #endregion

        #region 码表清零
        /// <summary>
        ///  码表清零
        /// </summary>
        public void Zeroing()
        {
            if (null == _tbl)
            {
                return;
            }

            _tbl.Zero();
        }
        #endregion

        #region 释放资源
        /// <summary>
        /// 释放资源
        /// </summary>
        public void Dispose()
        {
            try
            {
                foreach (var d in _detections)
                {
                    if (null == d)
                    {
                        continue;
                    }

                    d.Dispose();
                }

                foreach (var camera in _cameras)
                {
                    camera.Dispose();
                }

                if (null != _tbl)
                {
                    _tbl.Dispose();
                }

                _runThread.Abort();

                #region TWB20210902
                OS.Spin.SdkLayer.SdkMyCode.Flaw_ClassifyRelease(ref _handle_h, ref _handle_v, ref _handle_proc);
                #endregion
                OS.Spin.Common.Camera.JaiCamera.GetInstance().Close();
            }
            catch (Exception ex)
            {
                LogisTrac.WriteInfoLog(string.Format("FlawDetectionBl Dispose:{0}", ex.Message));
            }
        }
        #endregion

        #endregion

        #region private

        #region 接收相机数据
        /// <summary>
        ///  接收相机数据
        /// </summary>
        /// <param name="cameraId">相机编号</param>
        /// <param name="img">拍照数据</param>
        private void RecivedImg(int cameraId, Modle.Sdk.StructInfos.DATA_IMAGE1 img)
        {
            _imageId = OS.Spin.Running.Cache.GetInstance().CacheImgId;
            // 开始执行
            _detections[cameraId].StartDetection(img);
            //_finishedCameraCount++;
        }
        #endregion

        private double _cMeter = -1;
        public void TimerRecived(int type, double meter)
        {
            if (type == 1)
            {
                CameraTrigger(meter);
            }

            if (type == 0 && !_cMeter.Equals(meter))
            {
                // ImageSavingBl.Getinstance().PutCopy(new MFileCopy(2, OS.Spin.Running.Cache.GetInstance().CacheImgId, _cameraCount));
                _cMeter = meter;
            }
        }
        #region 执行分类处理
        /// <summary>
        /// 执行分类处理
        /// </summary>
        private void Classifying()
        {
            if (_cameraCount <= 0)
            {
                return;
            }
            var id = OS.Spin.Common.Machine.CpuTool.SetCpuID(_cameraCount);
            //将当前线程绑定到指定的cpu核心上
            OS.Spin.Common.Machine.CpuTool.SetThreadAffinityMask(OS.Spin.Common.Machine.CpuTool.GetCurrentThread(), new UIntPtr(id));

            while (true)
            {
                var time = 0;
                while (time < _cameraCount)
                {
                    _semaphorePool.Wait(string.Format("{0}{1}", OS.Spin.Running.Infos.Names.S_DETECTION_OK, _classId));
                    time++;
                }
                try
                {
                    //// 手动验布时不做算法处理
                    //if (!OS.Spin.Running.Cache.GetInstance().AutoFlaw)
                    //{
                    //    //_canRecived = true;
                    //    continue;
                    //}
                    // 获取图像原图
                    var flaw = new MFlawInfo();
                    for (var i = 0; i < _cameraCount; i++)
                    {
                        flaw.AddMat(OS.Spin.Running.Cache.GetInstance().Mats[i].Clone());
                    }
                    current_meter = (int)OS.Spin.Running.Cache.GetInstance().CMeter;
                    // 记录当前处理的图片编号
                    _reallyImageId = _imageId;
                    // 释放拍照触发
                    // 统计需要分类处理的瑕疵快照个数
                    var classify_num = 0;
                    var cache = OS.Spin.Running.Cache.GetInstance();

                    for (var i = 0; i < _cameraCount; i++)
                    {
                        var rois = cache.GetFore_Result(i);
                        //var rts = rois;
                        if (null == rois)
                        {
                            continue;
                        }
                        classify_num += rois.Count();
                    }
                    //TWB20210902
                    var result = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(Modle.Sdk.StructInfos.Fore_Result)) * classify_num);

                    var fIn = Marshal.AllocHGlobal(Marshal.SizeOf(typeof(Modle.Sdk.StructInfos.Fore_Result)) * classify_num);
                    long ptr = fIn.ToInt64();
                    var fores = new Modle.Sdk.StructInfos.Fore_Result[classify_num];
                    //IntPtr RPtr = new IntPtr(ptr);
                    var classify_cnt = 0;
                    for (int m = 0; m < _cameraCount; m++)
                    {
                        var rois = cache.GetFore_Result(m);
                        if (null == rois)
                        {
                            continue;
                        }
                        foreach (var roi in rois)
                        {

                            fores[classify_cnt] = roi;
                            //Marshal.StructureToPtr(fores[classify_cnt], RPtr, false);

                            ptr += Marshal.SizeOf(typeof(Modle.Sdk.StructInfos.Fore_Result));
                            //LogisTrac.WriteLog(String.Format("[检测] 结果 image id:{0}  label:{1}  prob:{2} prob_sum:{3} width:{4}  height:{5} x:{6} y:{7}", fores[classify_cnt].img_id, fores[classify_cnt].label, fores[classify_cnt].prob, fores[classify_cnt].prob_sum, fores[classify_cnt].width, fores[classify_cnt].height, fores[classify_cnt].xCenter_cam, fores[classify_cnt].yCenter_cam));
                            classify_cnt++;
                        }
                    }
                    // 组装分类数据
                    //var resultSize = Marshal.SizeOf(typeof(Modle.Sdk.StructInfos.CTOOLS_Result)) * classify_num;
                    //var results = Marshal.AllocHGlobal(resultSize);
                    LogisTrac.WriteInfoLog(String.Format("[分类] 开始 ={0}======", _reallyImageId));

                    //int final_labelsize = 0;
                    #region TWB20210902
                    _findWall.m_meters = current_meter;
                    //_findWall.th_prob0 = OS.Spin.Running.Cache.GetInstance().FindWall;
                    //_findWall.th_prob1 = OS.Spin.Running.Cache.GetInstance().JdWall;
                    //_findWall.th_prob2 = OS.Spin.Running.Cache.GetInstance().MhWall;
                    _findWall.cloths_kind = OS.Spin.Running.Cache.GetInstance().Cloths_kind;
                    #endregion


                    //int gain = 0;
                    #region TWB20210902
                    int finalNum = 0;
                    var a = OS.Spin.SdkLayer.SdkMyCode.Flaw_Classification(flaw.Mats[0].Data, flaw.Mats[1].Data, flaw.Mats[2].Data, fores, classify_num, result, ref finalNum, ref _findWall, _handle_h, _handle_v, _handle_proc);
                    //Marshal.FreeHGlobal(ptrIn);
                    #endregion

                    _gainValue = _findWall.m_gain;

                    LogisTrac.WriteInfoLog(String.Format("[相机增益] ===[{0}] ={1}======", _reallyImageId, _gainValue));
                    List<Fore_Result> rts = new List<Fore_Result>();
                    // 解析返回结果
                    for (var i = 0; i < finalNum; i++)
                    {
                        var ptr1 = (result + i * Marshal.SizeOf(typeof(Modle.Sdk.StructInfos.Fore_Result)));

                        rts.Add((Modle.Sdk.StructInfos.Fore_Result)Marshal.PtrToStructure(ptr1, typeof(Modle.Sdk.StructInfos.Fore_Result)));
                        LogisTrac.WriteInfoLog(String.Format("[分类] 结果 image id:{0}  label:{1}  prob:{2} prob_sum:{3} width:{4}  height:{5} x:{6} y:{7}", rts[i].img_id, rts[i].label, rts[i].prob, rts[i].prob_sum, rts[i].width, rts[i].height, rts[i].xCenter_cam, rts[i].yCenter_cam));
                    }

                    //for (var i = 0; i < finalNum; i++)
                    //{
                    //    LogisTrac.WriteLog(String.Format("[分类] 结果 ={0}===[{1}]===[{2}]==[{3}]==[{4}]==[{5}==[{7}]==[{8}]", fores[i].img_id, fores[i].label, fores[i].prob, fores[i].is_valid, fores[i].xCenter_cam, fores[i].yCenter_cam, fores[i].yCenter_cam, fores[i].width, fores[i].height));
                    //}

                    //for (var i = 0; i < final_labelout.Count(); i++)
                    //{
                    //    final_labelout[i].is_valid = 1;
                    //    final_labelout[i].deduction = 4;
                    //}

                    #region TWB20210902 方法删除
                    // OS.Spin.SdkLayer.SdkMyCode.Detc_Postprocess(fores, classify_num, label_poolarray, ref pool_size, final_labelout, ref final_labelsize, current_meter);
                    #endregion

                    //LogisTrac.WriteInfoLog(String.Format("[分类] 结束 ={0}===[{1}]===", _reallyImageId, final_labelsize));

                    //_canRecived = true;

                    if (finalNum == 0)
                    {
                        continue;
                    }

                    var index = cache.FlawId;
                    var subIndex = 0;
                    //先抠图后画框 本次只抠图
                    for (var i = 0; i < finalNum; i++)
                    {
                        var fl = rts[i];

                        //if (fl.is_repeat == 1)
                        //{
                        //    continue;
                        //}
                        MFinalFlawCell fr = new MFinalFlawCell();
                        fr.FlawName = _lableNames[fl.label];
                        fr.Meter = OS.Spin.Running.Cache.GetInstance().CMeter;
                        //MFinalFlawCell fr = new MFinalFlawCell()
                        //{
                        //    FlawName = _lableNames[fl.label],
                        //    Meter = OS.Spin.Running.Cache.GetInstance().CMeter
                        //};

                        flaw.AddFinalFlaw(fr);

                        var xMeter = 0.5;
                        if (fl.img_id > 1)
                        {
                            xMeter += 1;
                        }

                        if (fl.img_id > 2)
                        {
                            xMeter += 1;
                        }
                        var s = new MFlawSnap
                        {
                            No = string.Format("{0}_{1}", index, subIndex),
                            FlawName = _lableNames[fl.label],
                            #region TWB20210902
                            CenterX = xMeter,
                            CenterY = fl.yCenter_cam / 1200 * 0.3 + OS.Spin.Running.Cache.GetInstance().CMeter
                            #endregion
                        };

                        subIndex++;

                        var fMat = (flaw.Mats[fl.img_id]);
                        var newFMat = (flaw.Mats[fl.img_id]).Clone();
                        int start_x = fl.xCenter_cam * RS_FACTOR - ROI_WIDTH / 2;
                        int start_y = fl.yCenter_cam * RS_FACTOR - ROI_HEIGHT / 2;

                        if (start_x < 0)
                            start_x = 0;
                        if (start_x + ROI_WIDTH > fMat.Width)
                            start_x = fMat.Width - ROI_WIDTH;
                        if (start_y < 0)
                            start_y = 0;
                        if (start_y + ROI_HEIGHT > fMat.Height)
                            start_y = fMat.Height - ROI_HEIGHT;
                        s.SnapMat = (new Mat(fMat.Clone(), new OpenCvSharp.Rect(start_x, start_y, ROI_WIDTH, ROI_HEIGHT)));
                        //Cv2.Rectangle(fMat, new Point(start_x, start_y), new Point(start_x + ROI_WIDTH, start_y + ROI_HEIGHT), new Scalar(5, 5, 5), 2);

                        //Cv2.Circle(fMat, new Point(fore.xCenter_cam * RS_FACTOR, fore.yCenter_cam * RS_FACTOR), 150, new Scalar(255, 255, 255), 5);
                        flaw.AddSnapFlaws(s);
                        // LogisTrac.WriteLog(String.Format("[最终] 结果 ={0}===[{1}]===[{2}]==[{3}]", fores[i].img_id, fl.label, fl.prob, fl.is_valid));
                        //foreach (var fore in fores)

                    }
                    //本次循环只画框
                    for (var i = 0; i < finalNum; i++)
                    {
                        var fl = rts[i];
                        var xMeter = 0.5;
                        if (fl.img_id > 1)
                        {
                            xMeter += 1;
                        }
                        if (fl.img_id > 2)
                        {
                            xMeter += 1;
                        }
                        subIndex++;
                        var fMat = (flaw.Mats[fl.img_id]);
                        var newFMat = (flaw.Mats[fl.img_id]).Clone();
                        int start_x = fl.xCenter_cam * RS_FACTOR - ROI_WIDTH / 2;
                        int start_y = fl.yCenter_cam * RS_FACTOR - ROI_HEIGHT / 2;

                        if (start_x < 0)
                            start_x = 0;
                        if (start_x + ROI_WIDTH > fMat.Width)
                            start_x = fMat.Width - ROI_WIDTH;
                        if (start_y < 0)
                            start_y = 0;
                        if (start_y + ROI_HEIGHT > fMat.Height)
                            start_y = fMat.Height - ROI_HEIGHT;

                        //s.SnapMat = (new Mat(fMat, new OpenCvSharp.Rect(start_x, start_y, ROI_WIDTH, ROI_HEIGHT)));
                        Cv2.Rectangle(fMat, new Point(start_x, start_y), new Point(start_x + ROI_WIDTH, start_y + ROI_HEIGHT), new Scalar(5, 5, 5), 2);
                        //Cv2.Circle(fMat, new Point(fore.xCenter_cam * RS_FACTOR, fore.yCenter_cam * RS_FACTOR), 150, new Scalar(255, 255, 255), 5);
                        // LogisTrac.WriteLog(String.Format("[最终] 结果 ={0}===[{1}]===[{2}]==[{3}]", fores[i].img_id, fl.label, fl.prob, fl.is_valid));
                        //foreach (var fore in fores)
                    }

                    // 有用瑕疵存在时
                    if (flaw.FinalFlaws.Count > 0)
                    {
                        if (OS.Spin.Running.Cache.GetInstance().AutoFlaw)
                        {
                            TlWorker?.Invoke(11);
                        }
                        // 保存带瑕疵图片
                        ImageSavingBl.Getinstance().PutCopy(new MFileCopy(0, _reallyImageId, _cameraCount));
                        LogisTrac.WriteInfoLog(String.Format("PutCopy ={0}======", _reallyImageId));
                        DoAddFlawShot?.Invoke(flaw);
                        LogisTrac.WriteInfoLog(String.Format("DoAddFlawShot ={0}======", _reallyImageId));
                    }
                }
                catch (Exception ex)
                {
                    LogisTrac.WriteInfoLog("Classifying");
                    LogisTrac.WriteLog(ex);
                }
                finally
                {
                    _canRecived = true;
                    LogisTrac.WriteInfoLog(String.Format("正常结束 ={0}======", _reallyImageId));
                }
            }
        }
        #endregion


        /// <summary>
        /// 改变布匹参数
        /// </summary>
        /// <param name="kind"></param>
        public void UploadClothsKind(int kind)
        {
            try
            {
                foreach (var dl in _detections)
                {
                    dl.DoUploadClothsKind(kind);
                }
                _findWall.cloths_kind = kind;
                OS.Spin.Running.Cache.GetInstance().Cloths_kind = kind;
            }
            catch (Exception ex)
            {
                LogisTrac.WriteInfoLog(String.Format("UploadClothsKind: {0}", ex.Message));
            }
        }


        #region 获取三通道照片信息
        /// <summary>
        /// 获取三通道照片信息
        /// </summary>
        /// <param name="cammeraId"></param>
        /// <returns></returns>
        private Mat DoGetThreeChannelsMat(int cammeraId)
        {
            return _cameras[cammeraId].ThreeChannelMat;
        }
        #endregion

        #endregion

        #region ===== WARNING FINISHED ON 2019/06/06 BY Ding Shujie VERSION: V1.0.0 ======
        #endregion
    }
}