refactor(presenter): 移除重复的debug数据保存逻辑并优化点云图像生成。使用PointCloudCanvas重构棒材检测结果的可视化逻辑
This commit is contained in:
杰仔
parent
85697b1fd4
commit
b0c165fa8a
@ -51,15 +51,8 @@ int DetectPresenter::DetectHoles(
|
||||
return ERR_CODE(DEV_DATA_INVALID);
|
||||
}
|
||||
|
||||
// 保存debug数据
|
||||
// 调试模式下先保存原始激光线数据，便于后续离线复现问题。
|
||||
// debug保存点云已由 BasePresenter::DetectTask() 统一处理，此处不再重复保存
|
||||
std::string timeStamp = CVrDateUtils::GetNowTime();
|
||||
if (debugParam.enableDebug && debugParam.savePointCloud) {
|
||||
LOG_INFO("[Algo Thread] Debug mode is enabled, saving point cloud data\n");
|
||||
|
||||
std::string fileName = debugParam.debugOutputPath + "/Laserline_" + std::to_string(cameraIndex) + "_" + timeStamp + ".txt";
|
||||
dataLoader.SaveLaserScanData(fileName, laserLines, laserLines.size(), 0.0, 0, 0);
|
||||
}
|
||||
|
||||
int nRet = SUCCESS;
|
||||
|
||||
|
||||
@ -3,11 +3,7 @@
|
||||
#include "AlgoParamConverter.h"
|
||||
#include "IHandEyeCalib.h"
|
||||
#include <fstream>
|
||||
#define _USE_MATH_DEFINES
|
||||
#include <cmath>
|
||||
#include <memory>
|
||||
#include <QPainter>
|
||||
#include <QPen>
|
||||
#include <QColor>
|
||||
|
||||
namespace
|
||||
@ -68,17 +64,8 @@ int DetectPresenter::DetectRod(
|
||||
cameraCalibParam = &cameraCalibParamValue;
|
||||
}
|
||||
|
||||
// 保存debug数据
|
||||
// debug保存点云已由 BasePresenter::DetectTask() 统一处理，此处不再重复保存
|
||||
std::string timeStamp = CVrDateUtils::GetNowTime();
|
||||
if(debugParam.enableDebug && debugParam.savePointCloud){
|
||||
LOG_INFO("[Algo Thread] Debug mode is enabled, saving point cloud data\n");
|
||||
|
||||
// 获取当前时间戳，格式为YYYYMMDDHHMMSS
|
||||
std::string fileName = debugParam.debugOutputPath + "/Laserline_" + std::to_string(cameraIndex) + "_" + timeStamp + ".txt";
|
||||
|
||||
// 直接使用统一格式保存数据
|
||||
dataLoader.SaveLaserScanData(fileName, laserLines, laserLines.size(), 0.0, 0, 0);
|
||||
}
|
||||
|
||||
int nRet = SUCCESS;
|
||||
|
||||
@ -136,96 +123,32 @@ int DetectPresenter::DetectRod(
|
||||
const HECCalibResult calibResult = HECCalibResult::fromHomogeneousArray(clibMatrix);
|
||||
const HECEulerOrder hecEulerOrder = ToHandEyeEulerOrder(eulerOrder);
|
||||
|
||||
// 构建检测结果：生成点云图像
|
||||
// 1. 获取所有棒材的中心点用于可视化
|
||||
std::vector<std::vector<SVzNL3DPoint>> objOps;
|
||||
std::vector<SVzNL3DPoint> rodCenters;
|
||||
// 使用 PointCloudCanvas 生成点云图像（灰色底图 + 棒材中心/方向线标记）
|
||||
{
|
||||
PointCloudCanvas canvas = PointCloudCanvas::Create(xyzData);
|
||||
|
||||
for (const auto& rod : rodInfo) {
|
||||
SVzNL3DPoint pt;
|
||||
pt.x = rod.center.x;
|
||||
pt.y = rod.center.y;
|
||||
pt.z = rod.center.z;
|
||||
rodCenters.push_back(pt);
|
||||
}
|
||||
if (!rodCenters.empty()) {
|
||||
objOps.push_back(rodCenters);
|
||||
}
|
||||
for (size_t i = 0; i < rodInfo.size(); i++) {
|
||||
const auto& rod = rodInfo[i];
|
||||
|
||||
// 从点云数据生成投影图像（10cm边界），同时获取点云原始范围
|
||||
double margin = 100.0; // 10cm = 100mm
|
||||
double xMin, xMax, yMin, yMax;
|
||||
detectionResult.image = PointCloudImageUtils::GeneratePointCloudRetPointImage(xyzData, objOps, margin, &xMin, &xMax, &yMin, &yMax);
|
||||
// 绘制棒材中心点（红色）
|
||||
canvas.drawPoint(rod.center.x, rod.center.y, QColor(255, 0, 0), 6);
|
||||
|
||||
// 在图像上绘制棒材的轴向方向线
|
||||
if (!detectionResult.image.isNull() && !rodInfo.empty()) {
|
||||
QPainter painter(&detectionResult.image);
|
||||
painter.setRenderHint(QPainter::Antialiasing);
|
||||
|
||||
// 扩展边界（与GeneratePointCloudRetPointImage相同）
|
||||
xMin -= margin;
|
||||
xMax += margin;
|
||||
yMin -= margin;
|
||||
yMax += margin;
|
||||
|
||||
// 使用与GeneratePointCloudRetPointImage相同的参数
|
||||
int imgRows = detectionResult.image.height();
|
||||
int imgCols = detectionResult.image.width();
|
||||
int x_skip = 50;
|
||||
int y_skip = 50;
|
||||
|
||||
// 计算投影比例（与PointCloudImageUtils相同的方式）
|
||||
double y_rows = (double)(imgRows - y_skip * 2);
|
||||
double x_cols = (double)(imgCols - x_skip * 2);
|
||||
double x_scale = (xMax - xMin) / x_cols;
|
||||
double y_scale = (yMax - yMin) / y_rows;
|
||||
|
||||
// 使用统一的比例尺
|
||||
double scale = std::max(x_scale, y_scale);
|
||||
x_scale = scale;
|
||||
y_scale = scale;
|
||||
|
||||
// 计算点云在图像中居中的偏移量（与PointCloudImageUtils一致）
|
||||
double cloudWidth = (xMax - xMin) / scale;
|
||||
double cloudHeight = (yMax - yMin) / scale;
|
||||
int x_offset = x_skip + (int)((x_cols - cloudWidth) / 2);
|
||||
int y_offset = y_skip + (int)((y_rows - cloudHeight) / 2);
|
||||
|
||||
// 转换3D坐标到图像坐标的lambda函数（使用居中偏移）
|
||||
auto toImageCoord = [&](const SVzNL3DPoint& pt) -> QPointF {
|
||||
int px = (int)((pt.x - xMin) / x_scale + x_offset);
|
||||
int py = (int)((pt.y - yMin) / y_scale + y_offset);
|
||||
return QPointF(px, py);
|
||||
};
|
||||
|
||||
// 绘制棒材的轴向方向线
|
||||
for (const auto& rod : rodInfo) {
|
||||
// 绘制轴向方向线（红色）
|
||||
double len = 60;
|
||||
QPen axisPen(QColor(255, 0, 0), 2);
|
||||
painter.setPen(axisPen);
|
||||
|
||||
SVzNL3DPoint pt0 = { rod.center.x - len * rod.axialDir.x,
|
||||
rod.center.y - len * rod.axialDir.y,
|
||||
rod.center.z - len * rod.axialDir.z };
|
||||
SVzNL3DPoint pt1 = { rod.center.x + len * rod.axialDir.x,
|
||||
rod.center.y + len * rod.axialDir.y,
|
||||
rod.center.z + len * rod.axialDir.z };
|
||||
|
||||
QPointF imgPt0 = toImageCoord(pt0);
|
||||
QPointF imgPt1 = toImageCoord(pt1);
|
||||
painter.drawLine(imgPt0, imgPt1);
|
||||
double ax0 = rod.center.x - len * rod.axialDir.x;
|
||||
double ay0 = rod.center.y - len * rod.axialDir.y;
|
||||
double ax1 = rod.center.x + len * rod.axialDir.x;
|
||||
double ay1 = rod.center.y + len * rod.axialDir.y;
|
||||
canvas.drawLine(ax0, ay0, ax1, ay1, QColor(255, 0, 0), 2);
|
||||
|
||||
// 绘制起点到终点线段（绿色）
|
||||
QPen segPen(QColor(0, 255, 0), 2);
|
||||
painter.setPen(segPen);
|
||||
canvas.drawLine(rod.startPt.x, rod.startPt.y, rod.endPt.x, rod.endPt.y, QColor(0, 255, 0), 2);
|
||||
|
||||
SVzNL3DPoint startPt = { rod.startPt.x, rod.startPt.y, rod.startPt.z };
|
||||
SVzNL3DPoint endPt = { rod.endPt.x, rod.endPt.y, rod.endPt.z };
|
||||
QPointF imgStart = toImageCoord(startPt);
|
||||
QPointF imgEnd = toImageCoord(endPt);
|
||||
painter.drawLine(imgStart, imgEnd);
|
||||
// 中心点白色编号
|
||||
canvas.drawText(rod.center.x, rod.center.y, QString("%1").arg(i + 1), Qt::white, 14);
|
||||
}
|
||||
|
||||
detectionResult.image = canvas.image();
|
||||
}
|
||||
|
||||
// 转换检测结果为UI显示格式（使用机械臂坐标系数据）
|
||||
|
||||
@ -45,17 +45,8 @@ int DetectPresenter::DetectWorkpieceHole(
|
||||
cameraCalibParam = &cameraCalibParamValue;
|
||||
}
|
||||
|
||||
// 保存debug数据
|
||||
// debug保存点云已由 BasePresenter::DetectTask() 统一处理，此处不再重复保存
|
||||
std::string timeStamp = CVrDateUtils::GetNowTime();
|
||||
if(debugParam.enableDebug && debugParam.savePointCloud){
|
||||
LOG_INFO("[Algo Thread] Debug mode is enabled, saving point cloud data\n");
|
||||
|
||||
// 获取当前时间戳，格式为YYYYMMDDHHMMSS
|
||||
std::string fileName = debugParam.debugOutputPath + "/Laserline_" + std::to_string(cameraIndex) + "_" + timeStamp + ".txt";
|
||||
|
||||
// 直接使用统一格式保存数据
|
||||
dataLoader.SaveLaserScanData(fileName, laserLines, laserLines.size(), 0.0, 0, 0);
|
||||
}
|
||||
|
||||
int nRet = SUCCESS;
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user