装载MAT文件与地面真相数据。装载的“caltech_cordova1_laneAndVehicleGroundTruth.mat”）;

创建一个VideoReader对象来读取视频的帧。videoName = loaded.videoName;fileReader = videereader (videoName);基本真相数据以时间表的形式组织。。gtdata = loaded.groundTruthData;显示地面真相数据的前几行。。头(gtdata)

阅读视频的第一帧。frame = readFrame(fileReader);提取第一帧中的所有车道点。lanePoints = gtdata. laneborders {1};在第一帧提取车辆的包围盒。= gtdata.Vehicles{1};叠加右车道点和车辆边界框。帧= insertMarker(frame, lanePoints{2}，“X”）;帧= insertObjectAnnotation(帧，“矩形”vehicleBBox,“汽车”）;显示第一帧的地面真实数据。图imshow(框架)

设置monoSensorHelper来处理视频。monocamersensor = loaded.sensor;monoSensorHelper = helperMonoSensor(moncamersensor);为测量创建具有相同时间向量的新时间表。。measurements =时间表(gtdata.Time);设置时间表列，以保存车道边界和车辆数据。numFrames = floor(fileReader.FrameRate*fileReader.Duration);测量。laneborders = cell(numFrames, 2);测量。= cell(numFrames, 1);gtdata。LanesInVehicleCoord = cell(numFrames, 2);倒回视频到t = 0，并创建一个帧索引来保持当前%框架。fileReader。CurrentTime = 0;frameIndex = 0;循环视频文件，直到没有新的帧。而hasFrame(fileReader) frameIndex = frameIndex+1;frame = readFrame(fileReader);使用processFrame方法计算检测。此方法可以替换为自定义车道检测方法。。检测= processFrame(monoSensorHelper, frame);存储估计的车道边界和车辆检测。。测量。laneborders {frameIndex} = [tests . leftegoboundary ..．.detections.rightEgoBoundary];测量。VehicleDetections{frameIndex} = tests . vehicleboxes;为便于比较，将地面真道点转换为%车辆坐标系统。gtPointsThisFrame = gtdata. laneborders {frameIndex};vehiclePoints = cell(1, numel(gtPointsThisFrame));为ii = 1:numel(gtPointsThisFrame) vehiclePoints{ii} = imageToVehicle(monocamersensor, gtPointsThisFrame{ii});结束存储以车辆坐标表示的地面真值点。gtdata。LanesInVehicleCoord{frameIndex} = vehiclePoints;结束

倒回视频到t = 0。fileReader。CurrentTime = 0;阅读视频的第一帧。frame = readFrame(fileReader);birdsEyeImage = transformImage(monoSensorHelper.)BirdsEyeConfig,框架);在鸟瞰图中提取第一帧的右车道点。。firstFrameVehiclePoints = gtdata.LanesInVehicleCoord{1};pointsInBEV = vehicleToImage(monoSensorHelper.)BirdsEyeConfig, firstFrameVehiclePoints {2});在框架上叠加点。插入标记(birdsEyeImage, pointsInBEV，“X”，“大小”6);%在鸟瞰视图中显示转换的点。图imshow (birdsEyeImage)

阈值= 0.25;米%[numMatches, numMisses, numfalse positive, assignments] =.．.evaluateLaneBoundaries(测量。LaneBoundaries,.．.gtdata。LanesInVehicleCoord,.．.阈值);disp ([“比赛次数:”num2str (numMatches)]);disp ([“失误次数:”num2str (numMisses)]);disp ([“误报数量:”num2str (numFalsePositives)]);

averageDistance = helperComputeLaneStatistics(测量。LaneBoundaries,.．.gtdata。LanesInVehicleCoord,.．.作业,@mean);绘制估算值和实际情况之间的平均距离。。图阻止(gtdata。时间,averageDistance) title(“估算值与实际情况之间的平均距离”网格)在ylabel (“以米为单位的距离”)传说(的左边界，“右边界”)

使用monoCamera计算估计的点位置。[estVehiclePoints, estImagePoints] = helperget对应点(monocamersensor，.．.测量。LaneBoundaries,.．.gtdata。LanesInVehicleCoord,.．.作业);将估计的车道添加到测量时间表。。测量。EstimatedLanes = estImagePoints;测量。LanesInVehicleCoord = estVehiclePoints;用可视化所需的所有变量创建一个新的时间表。。Names = {“LanePoints”；“DetectedLanePoints”};类型= labelType({“行”；“行”});标签defs =表(名称，类型，“VariableNames”, {“名字”，“类型”});visualizeInFrame = schedule (gtdata. visualizeInFrame)时间,.．.gtdata。LaneBoundaries,.．.测量。EstimatedLanes,.．.“VariableNames”、名称);创建groundTruth对象。数据源= groundTruthDataSource(videoName);datatovisualization = groundTruth(数据源，labelDefs, visualizeInFrame);将前一节的所有结果保存在distanceData中。垫子临时文件夹。dataToLoad = [tempdir .“distanceData.mat”];保存(dataToLoad,“monoSensorHelper”，“videoName”，“测量”，“gtdata”，“averageDistance”）;

函数stat = helperComputeLaneStatistics(estModels, gtPoints, assignments, fcnHandle) numFrames = length(estModels);默认情况下使左右估计NaN表示缺乏%的数据。stat = NaN*ones(numFrames, 2);为frameInd = 1:numFrames%默认使左右估计NaN。stat(frameInd，:) = NaN*ones(2,1);为idx = 1:length(estModels{frameInd})忽略假阳性作业。如果分配{frameInd}(idx) == 0继续；结束estModelInFrame中的第k个边界与第k个边界匹配元素在gtPointsInFrame中被赋值索引。thisModel = estModels{frameInd}(idx);thisGT = getpoints {frameInd}{assignments{frameInd}(idx)};thisGT model = driving. inner . piecewiselinearboundary (thisGT);如果mean(thisGTModel.Points(:，2)) > 0左车道%xPoints = thisGTModel.Points(:，1);yDist = 0 (size(xPoints));为index = 1:numel(xPoints) gtYPoints = thisGTModel.computeBoundaryModel(xPoints(index));testYPoints = thisModel.computeBoundaryModel(xPoints(index));yDist(index) = abs(testYPoints-gtYPoints);结束stat(frameInd, 1) = fcnHandle(yDist);其他的右车道%xPoints = thisGTModel.Points(:，1);yDist = 0 (size(xPoints));为index = 1:numel(xPoints) gtYPoints = thisGTModel.computeBoundaryModel(xPoints(index));testYPoints = thisModel.computeBoundaryModel(xPoints(index));yDist(index) = abs(testYPoints-gtYPoints);结束stat(frameInd, 2) = fcnHandle(yDist);结束结束结束结束

函数[vehiclePoints, imagePoints] = helperget对应点(monocamersensor, estModels, gtPoints, assignments) numFrames = length(estModels);imagePoints = cell(numFrames, 1);= cell(numFrames, 1);为frameInd = 1:numFrames如果isempty(assignments{frameInd}) imagePointsInFrame = [];vehiclePointsInFrame = [];其他的estModelInFrame = estModels{frameInd};gtPointsInFrame = gtPoints{frameInd};imagePointsInFrame = cell(length(estModelInFrame)， 1);= cell(length(estModelInFrame)， 1);为idx = 1:length(estModelInFrame)忽略假阳性作业。如果分配{frameInd}(idx) == 0 imagePointsInFrame{idx} = [NaN NaN];继续；结束estModelInFrame中的第k个边界与第k个边界匹配元素在gtPointsInFrame中被赋值索引。thisModel = estModelInFrame(idx);thisGT = gtPointsInFrame{assignments{frameInd}(idx)};xPoints = thisGT(:， 1);yPoints = thisModel.computeBoundaryModel(xPoints);vehiclePointsInFrame{idx} = [xPoints, yPoints];imagePointsInFrame{idx} = vehicleToImage(monocamersensor， [xPoints yPoints]);结束结束vehiclePoints{frameInd} = vehiclePointsInFrame;imagePoints{frameInd} = imagePointsInFrame;使用imagePoints[]而不是{}来遵从groundTruth对象。如果isempty(imagePoints{frameInd}) imagePoints{frameInd} = [];结束如果isempty(vehiclePoints{frameInd}) vehiclePoints{frameInd} = [];结束结束结束