MvvmScaffoldFrame48/MvvmScaffoldFrame48.DLL/AlgorithmProcess/ProcessingAlgorithm_CCDShuLi.cs

using MvCameraControl;
using MvvmScaffoldFrame48.DLL.ConfigTools;
using MvvmScaffoldFrame48.DLL.LogTools;
using MvvmScaffoldFrame48.DLL.SystemTools;
using MvvmScaffoldFrame48.DLL.ThreadManager;
using MvvmScaffoldFrame48.Model.StorageModel.ImageAlgorithm.ShuLI;
using MvvmScaffoldFrame48.Model.StorageModel.Configs;
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Drawing;
using System.Linq;
using System.Text;
using System.Threading;
using System.Threading.Tasks;
using MvvmScaffoldFrame48.DLL.ImageAlgorithm;

namespace MvvmScaffoldFrame48.DLL.AlgorithmProcess
{
    public class ProcessingAlgorithm_CCDShuLi : IImageProcessingAlgorithmHikVision
    {
        private List<ActiveObjectClassModel> activeObjects = new List<ActiveObjectClassModel>(); // 当前跟踪中的物体
        private List<ActiveObjectClassModel> lostObjects = new List<ActiveObjectClassModel>();
        private List<ActiveObjectClassModel> historyActiveObjects = new List<ActiveObjectClassModel>(); // 历史物体
        BoundRectangleClass BoundRectangle = new BoundRectangleClass();
        private long currentLine = 0; //行数记录
        private int ObjectNum = 0;
        private int ChannelWidth = 0;//每个区域的宽度
        public int HistoryNum { get { return _HistoryNum; } }
        private int _HistoryNum = 0;
        public int HistoryOkNum {  get { return _HistoryOkNum; } }
        private int _HistoryOkNum = 0;
        public int HistoryNgNum {  get { return _HistoryNgNum; } }
        private int _HistoryNgNum = 0;
        private ShuLiConfigClassModel shuLiConfig = new ShuLiConfigClassModel();
        public List<int> ChannelsRoi { get { return _ChannelsRoi; } }
        private List<int> _ChannelsRoi = new List<int>();
        private bool IsPrintLightOnError = false;
        public string AlgorithmName => "ProcessingAlgorithm_CCDShuLi";

        public void Configure(string parameters)
        {
            var parameter = XMLReadWrite.DeserializeFromString<ShuLiConfigClassModel>(parameters);
            if (parameter != null)
            {
                shuLiConfig = parameter;
            }
        }

        public object GetParameters()
        {
            return shuLiConfig;
        }

        public string GetSaveJson()
        {
            return XMLReadWrite.SerializeToString(shuLiConfig);
        }

        public object ProcessImage(IFrameOut imageData, int cameraId)
        {
            bool result = ProcessImageSequence(imageData, out List<ActiveObjectClassModel> resultValue);
            if (result)
            {
                return resultValue;
            }
            else
            {
                return null;
            }
        }

        /// <summary>
        /// 处理图像序列的主入口
        /// </summary>
        /// <param name="image">图像像素数据</param>
        /// <param name="ImageWidth">图像宽</param>
        /// <param name="currentLine">当前行数</param>
        /// <returns>检测到的物体总数</returns>
        public bool ProcessImageSequence(IFrameOut image,out List<ActiveObjectClassModel> resultValue)
        {
            bool result = false;
            for (int i = 0; i < image.Image.Height; i++)
            {
                ProcessLine(image, i);
                currentLine += 1;
            }
            //识别到结果并输出
            // 清理超时未更新的物体
            lostObjects = activeObjects
                .Where(o => (currentLine - o.LastSeenLine) > shuLiConfig.MAX_GAP || (o.LastSeenLine - o.StartLine) > shuLiConfig.MAX_Idetify_Height)
                .ToList();
            lostObjects.ForEach(o => activeObjects.Remove(o));

            resultValue = new List<ActiveObjectClassModel>();

            // 有物体转变为活跃物体，返回值转为true
            if (lostObjects.Count > 0)
            {
                foreach (var item in lostObjects)
                {
                    //噪点判定
                    if (item.Area < shuLiConfig.NoiseFilter_Threshold)
                    {
                        item.StateCode = 9;
                        //LOG.log(string.Format("噪点过滤，噪点面积：{0}", item.Area), 6);
                        continue;
                    }
                    //转为历史物体，添加缺少的参数
                    item.Num = Interlocked.Increment(ref ObjectNum);
                    //item.ChannelNO = ActiveChannel(item);
                    item.EndCheckTime = DateTime.Now;
                    item.MaxLength = GetActionMaxLength(item.RowsData);
                    item.PictureEndReadTime = TimeStampTools.FromUnixTimestamp((long)image.HostTimeStamp);
                    var QutuYanshiTime = (DateTime.Now - item.PictureEndReadTime).TotalMilliseconds;
                    if (QutuYanshiTime > 30)
                    {
                        Console.WriteLine($"结果输出延时超过了30ms，耗时{QutuYanshiTime}");
                    }
                    if ((item.LastSeenLine - item.StartLine) > shuLiConfig.MAX_Idetify_Height)
                    {
                        item.StateCode = 7;
                    }
                    else if (shuLiConfig.PandingCode != -1)
                    {
                        if (item.StateCode != -1)
                        {
                            if (item.StateCode == 8)
                            {
                                //TxtLog.log(string.Format("颗粒编号{0}：疑似叠片或缺损", item.Num));
                            }
                        }
                        else if (item.Area < shuLiConfig.MinArea
                            && (shuLiConfig.PandingCode == 2 || shuLiConfig.PandingCode == 1))
                        {
                            item.StateCode = 5;
                        }
                        else if (item.Area > shuLiConfig.MaxArea
                            && (shuLiConfig.PandingCode == 2 || shuLiConfig.PandingCode == 1))
                        {
                            item.StateCode = 6;
                        }
                        else if (item.MaxLength < shuLiConfig.MIN_Object_LENGTH
                            && (shuLiConfig.PandingCode == 2 || shuLiConfig.PandingCode == 0))
                        {
                            item.StateCode = 2;
                        }
                        else if (item.MaxLength > shuLiConfig.MAX_Object_LENGTH
                            && (shuLiConfig.PandingCode == 2 || shuLiConfig.PandingCode == 0))
                        {
                            item.StateCode = 1;
                        }
                        else
                        {
                            item.StateCode = 0;
                        }
                    }
                    resultValue.Add(item);
                }
                if (resultValue.Count > 0)
                {
                    result = true;
                }
            }
            else
            {
                resultValue.Clear();
            }

            // 累加到总数并从活跃物体转移到历史物体
            resultValue.ForEach(o => TryAdd(historyActiveObjects, o, 2500));

            return result;
        }

        /// <summary>
        /// 处理单行像素数据
        /// 返回值为false的时候无活跃物体转变为历史物体
        /// 返回值为true的时候有活跃物体转变为历史物体
        /// </summary>
        /// <param name="image">当前行像素数组</param>
        private bool ProcessLine(IFrameOut imagedata, int RowNo)
        {
            Stopwatch stopwatch = Stopwatch.StartNew();
            bool result = false;
            // 步骤1：检测当前行的有效区域
            var currentRegions = FindValidRegions(imagedata.Image, RowNo);

            if (currentRegions.Count == 1)
            {
                if (currentRegions[0].End - (currentRegions[0]).Start + 1 == imagedata.Image.Width)
                {
                    if (!IsPrintLightOnError)
                    {
                        //FaultLog.RecordLogMessage("当前行有效区域为整行,检查视野和光源", 5);
                        IsPrintLightOnError = true;
                    }
                    return false;
                }
            }
            IsPrintLightOnError = false;
            stopwatch.Stop();
            if (stopwatch.ElapsedMilliseconds > 1)
            {
                //FaultLog.RecordErrorMessage($"ShuLiClass-ProcessLine:图像连通域检测超时，此次识别耗时：{stopwatch.Elapsed}");
            }
            stopwatch.Restart();
            //lock (_lockObj)
            //{
                foreach (var region in currentRegions)
                {
                    // 查找全部可合并的活跃物体（有重叠+在允许间隔内）
                    var matcheds = activeObjects.Where(o =>
                        IsOverlapping(o, region) &&
                        (currentLine - o.LastSeenLine - 1) <= shuLiConfig.MAX_GAP).ToList();
                    //当有多个可合并的活跃物体时，将多个物体合并
                    if (matcheds.Count >= 2)
                    {
                        // 合并有效区域队列
                        var CopeRowsData = new List<RowStartEndColModel>();
                        matcheds.ForEach(o => CopeRowsData = CopeRowsData.Concat(o.RowsData).ToList());
                        // 合并有效区域并保存在新的区域中
                        var MergeMatched = new ActiveObjectClassModel
                        {
                            MinStartCol = matcheds.Min(o => o.MinStartCol),
                            MaxEndCol = matcheds.Max(o => o.MaxEndCol),
                            StartLine = matcheds.Min(o => o.StartLine),
                            LastSeenLine = matcheds.Max(o => o.LastSeenLine),
                            LastSeenLineStartCol = matcheds.Min(o => o.LastSeenLineStartCol),
                            LastSeenLineEndCol = matcheds.Max(o => o.LastSeenLineEndCol),
                            StartCheckTime = matcheds.Min(o => o.StartCheckTime),
                            EndCheckTime = matcheds.Max(o => o.EndCheckTime),
                            PictureStartReadTime = matcheds.Min(o=>o.PictureStartReadTime),
                            Area = matcheds.Sum(o => o.Area),
                            RowsData = CopeRowsData,
                            ImageWidth = matcheds.FirstOrDefault().ImageWidth,
                            //StateCode = 8
                        };
                        // 从活跃区域中删除被合并的区域
                        matcheds.ForEach(o => activeObjects.Remove(o));
                        // 保存新的区域到活跃区域中
                        activeObjects.Add(MergeMatched);
                    }

                    // 搜获可用且可合并的活跃区域
                    var matched = activeObjects.FirstOrDefault(o =>
                        IsOverlapping(o, region) &&
                        (currentLine - o.LastSeenLine - 1) <= shuLiConfig.MAX_GAP);
                    if (matched != null)
                    {
                        // 合并区域：扩展物体边界并更新状态
                        matched.MinStartCol = Math.Min(matched.MinStartCol, region.Start);
                        matched.MaxEndCol = Math.Max(matched.MaxEndCol, region.End);
                        matched.Area += region.End - region.Start + 1;
                        matched.LastSeenLine = currentLine;
                        matched.RowsData.Add(new RowStartEndColModel
                        {
                            StartCol = region.Start,
                            EndCol = region.End,
                            RowsCol = currentLine,
                        });
                        matched.LastSeenLineStartCol = region.Start;
                        matched.LastSeenLineEndCol = region.End;
                    }
                    else
                    {
                    // 创建新物体（首次出现的区域）
                    activeObjects.Add(new ActiveObjectClassModel
                    {
                        MinStartCol = region.Start,
                        MaxEndCol = region.End,
                        StartLine = currentLine,
                        LastSeenLine = currentLine,
                        LastSeenLineStartCol = region.Start,
                        LastSeenLineEndCol = region.End,
                        StartCheckTime = DateTime.Now,
                        PictureStartReadTime = TimeStampTools.FromUnixTimestamp((long)imagedata.HostTimeStamp),
                        Area = region.End - region.Start + 1,
                        ImageWidth = shuLiConfig.ImageWidth,
                        RowsData = new List<RowStartEndColModel> {
                                new RowStartEndColModel {
                                    StartCol = region.Start,
                                    EndCol = region.End,
                                    RowsCol = currentLine,
                                }
                            }
                    });
                    }
                }
            //}
            stopwatch.Stop();
            if (stopwatch.ElapsedMilliseconds > 1)
            {
                //FaultLog.RecordErrorMessage($"ShuLiClass-ProcessLine:图像区域合并超时，此次识别耗时：{stopwatch.Elapsed}");
            }

            currentRegions.Clear();

            return result;
        }
        List<ValidRegionModel> regions = new List<ValidRegionModel>();
        /// <summary>
        /// 检测有效物体区域（横向连续黑色像素段）
        /// </summary>
        /// <param name="line">当前行像素数组</param>
        /// <returns>有效区域列表（起始/结束位置）</returns>
        private List<ValidRegionModel> FindValidRegions(IImage image, int RowNo)
        {
            regions.Clear();
            int start = -1; // 当前区域起始标记
            int end = -1;
            // 遍历所有像素列
            if (shuLiConfig.IsIdentifyRoiOpen)
            {
                for (int i = (int)image.Width * RowNo + shuLiConfig.IdentifyStartX; i < (int)image.Width * RowNo + shuLiConfig.IdentifyStopX; i++)
                {
                    if (image.PixelData[i] < shuLiConfig.RegionThreshold) // 发现黑色像素
                    {
                        if (start == -1) start = i % (int)image.Width; // 开始新区域
                    }
                    else if (start != -1) // 遇到白色像素且存在进行中的区域
                    {
                        end = (i - 1) % (int)image.Width;
                        // 检查区域宽度是否达标
                        if (end - start >= shuLiConfig.NoiseFilter_Threshold)
                        {
                            regions.Add(new ValidRegionModel()
                            {
                                Start = start,
                                End = end
                            }); // 记录有效区域
                        }
                        start = -1; // 重置区域标记
                        end = -1;
                    }
                }
            }
            else
            {
                for (int i = (int)image.Width * RowNo; i < (int)image.Width * (RowNo + 1); i++)
                {
                    if (image.PixelData[i] < shuLiConfig.RegionThreshold) // 发现黑色像素
                    {
                        if (start == -1) start = i % (int)image.Width; // 开始新区域
                    }
                    else if (start != -1) // 遇到白色像素且存在进行中的区域
                    {
                        end = (i - 1) % (int)image.Width;
                        // 检查区域宽度是否达标
                        if (end - start >= shuLiConfig.NoiseFilter_Threshold)
                        {
                            regions.Add(new ValidRegionModel()
                            {
                                Start = start,
                                End = end
                            }); // 记录有效区域
                        }
                        start = -1; // 重置区域标记
                        end = -1;
                    }
                }
            }


            // 处理行尾未闭合的区域
            if (start != -1 && image.Width - start >= shuLiConfig.NoiseFilter_Threshold)
            {
                regions.Add(new ValidRegionModel()
                {
                    Start = start,
                    End = (int)image.Width - 1
                });
            }
            return regions;
        }

        /// <summary>
        /// 判断区域重叠（与活跃物体的横向坐标重叠检测）
        /// </summary>
        /// <param name="obj">活跃物体</param>
        /// <param name="region">当前区域</param>
        /// <returns>是否发生重叠</returns>
        private bool IsOverlapping(ActiveObjectClassModel obj, ValidRegionModel region)
        {
            // 判断区域是否不相交的逆条件
            return !(region.End < obj.LastSeenLineStartCol || region.Start > obj.LastSeenLineEndCol);
        }

        /// <summary>
        /// 历史数据队列添加数据
        /// </summary>
        /// <param name="list">历史数据队列</param>
        /// <param name="item">添加的数据</param>
        /// <param name="maxSize">队列最大数据量</param>
        /// <returns></returns>
        private bool TryAdd(List<ActiveObjectClassModel> list, ActiveObjectClassModel item, int maxSize)
        {
            list.Add(item);
            Interlocked.Increment(ref _HistoryNum);
            if(item.StateCode == 0)
            {
                Interlocked.Increment(ref _HistoryOkNum);
            }
            else
            {
                Interlocked.Increment(ref _HistoryNgNum);
            }
            if (list.Count > maxSize)
            {
                list.Remove(list[list.Count - maxSize]);
            }
            return true;
        }

        /// <summary>
        /// 获取结果最长长边
        /// </summary>
        /// <param name="Rows"></param>
        /// <returns></returns>
        private double GetActionMaxLength(List<RowStartEndColModel> Rows)
        {
            //外轮廓点集
            List<Point> points = Rows.Select(o => new Point(o.StartCol, (int)o.RowsCol)).ToList();
            points.AddRange(Rows.Select(o => new Point(o.EndCol, (int)o.RowsCol)).ToList());
            //获取凸包点集
            points = BoundRectangle.ConvexHull(points);
            //获取凸包长度最长的外接矩形
            var result = BoundRectangle.RotatingCalipers(points);
            if (result != null)
            {
                return result.Height;
            }
            else
            {
                return 0;
            }
        }


        /// <summary>
        /// 通道区域判定
        /// </summary>
        /// <param name="activeObject"></param>
        /// <returns></returns>
        private int ActiveChannel(ActiveObjectClassModel activeObject)
        {
            int result = -1;
            int StartChannel = activeObject.MinStartCol / ChannelWidth;
            int EndChannel = activeObject.MaxEndCol / ChannelWidth;
            if (StartChannel == EndChannel)
            {
                result = StartChannel;
            }
            else if (EndChannel - StartChannel > 1)
            {
                Console.WriteLine("ActiveChannel-Error");
                //error
            }
            else
            {
                result = _ChannelsRoi[StartChannel] - activeObject.MinStartCol > activeObject.MaxEndCol - _ChannelsRoi[StartChannel] ? StartChannel : EndChannel;
            }
            return result;
        }
    }
}