目录
说明
百度网盘AI大赛-表格检测的第2名方案。
该算法包含表格边界框检测、表格分割和表格方向识别三个部分,首先,ppyoloe-plus-x 对边界框进行预测,并对置信度较高的表格边界框(box)进行裁剪。裁剪后的单个表格实例会送入到DBNet中进行语义分割,分割结果通过opencv轮廓处理获得表格关键点(point)。之后,我们根据DBNet计算的关键点在裁剪后的单个表格实例上绘制表格边界。最后,PP-LCNet结合表格边界先验和表格实例图像,对表格的方向进行预测,并根据之前定义的几何轮廓点与语义轮廓点的对应关系,将几何轮廓点映射为语义轮廓点。
本文使用C# OnnxRuntime 实现百度网盘AI大赛-表格检测第2名方案第二部分-表格分割
效果
模型
Model Properties
-------------------------
date:2024-10-28T08:16:43.725877
description:Ultralytics YOLO11l-seg model trained on coco-seg.yaml
author:Ultralytics
version:8.3.23
task:segment
license:AGPL-3.0 License (https://ultralytics.com/license)
docs:https://docs.ultralytics.com
stride:32
batch:1
imgsz:[800, 800]
names:{0: 'table'}
---------------------------------------------------------------
Inputs
-------------------------
name:images
tensor:Float[1, 3, 800, 800]
---------------------------------------------------------------
Outputs
-------------------------
name:output0
tensor:Float[1, 37, 13125]
name:output1
tensor:Float[1, 32, 200, 200]
---------------------------------------------------------------
项目
代码
using Microsoft.ML.OnnxRuntime;
using Microsoft.ML.OnnxRuntime.Tensors;
using OpenCvSharp;
using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Linq;
using System.Security.Cryptography;
using System.Text;
using System.Web;
using System.Windows.Forms;
using static System.Net.Mime.MediaTypeNames;
namespace Onnx_Yolov8_Demo
{
public partial class Form1 : Form
{
public Form1()
{
InitializeComponent();
}
string fileFilter = "*.*|*.bmp;*.jpg;*.jpeg;*.tiff;*.tiff;*.png";
string image_path = "";
string startupPath;
string classer_path;
DateTime dt1 = DateTime.Now;
DateTime dt2 = DateTime.Now;
string model_path;
Mat image;
SegmentationResult result_pro;
Mat result_image;
SessionOptions options;
InferenceSession onnx_session;
Tensor<float> input_tensor;
List<NamedOnnxValue> input_container;
IDisposableReadOnlyCollection<DisposableNamedOnnxValue> result_infer;
DisposableNamedOnnxValue[] results_onnxvalue;
Tensor<float> result_tensors_det;
Tensor<float> result_tensors_proto;
private void button1_Click(object sender, EventArgs e)
{
OpenFileDialog ofd = new OpenFileDialog();
ofd.Filter = fileFilter;
if (ofd.ShowDialog() != DialogResult.OK) return;
pictureBox1.Image = null;
image_path = ofd.FileName;
pictureBox1.Image = new Bitmap(image_path);
textBox1.Text = "";
image = new Mat(image_path);
pictureBox2.Image = null;
}
private void button2_Click(object sender, EventArgs e)
{
if (image_path == "")
{
return;
}
button2.Enabled = false;
System.Windows.Forms.Application.DoEvents();
// 配置图片数据
image = new Mat(image_path);
int max_image_length = image.Cols > image.Rows ? image.Cols : image.Rows;
Mat max_image = Mat.Zeros(new OpenCvSharp.Size(max_image_length, max_image_length), MatType.CV_8UC3);
Rect roi = new Rect(0, 0, image.Cols, image.Rows);
image.CopyTo(new Mat(max_image, roi));
float[] det_result_array = new float[37 * 13125];
float[] proto_result_array = new float[32 * 200 * 200];
float[] factors = new float[4];
factors[0] = factors[1] = (float)(max_image_length / 800.0);
factors[2] = image.Rows;
factors[3] = image.Cols;
// 将图片转为RGB通道
Mat image_rgb = new Mat();
Cv2.CvtColor(max_image, image_rgb, ColorConversionCodes.BGR2RGB);
Mat resize_image = new Mat();
Cv2.Resize(image_rgb, resize_image, new OpenCvSharp.Size(800, 800));
// 输入Tensor
for (int y = 0; y < resize_image.Height; y++)
{
for (int x = 0; x < resize_image.Width; x++)
{
input_tensor[0, 0, y, x] = resize_image.At<Vec3b>(y, x)[0] / 255f;
input_tensor[0, 1, y, x] = resize_image.At<Vec3b>(y, x)[1] / 255f;
input_tensor[0, 2, y, x] = resize_image.At<Vec3b>(y, x)[2] / 255f;
}
}
//将 input_tensor 放入一个输入参数的容器,并指定名称
input_container.Add(NamedOnnxValue.CreateFromTensor("images", input_tensor));
dt1 = DateTime.Now;
//运行 Inference 并获取结果
result_infer = onnx_session.Run(input_container);
dt2 = DateTime.Now;
// 将输出结果转为DisposableNamedOnnxValue数组
results_onnxvalue = result_infer.ToArray();
// 读取第一个节点输出并转为Tensor数据
result_tensors_det = results_onnxvalue[0].AsTensor<float>();
result_tensors_proto = results_onnxvalue[1].AsTensor<float>();
det_result_array = result_tensors_det.ToArray();
proto_result_array = result_tensors_proto.ToArray();
resize_image.Dispose();
image_rgb.Dispose();
result_pro = new SegmentationResult(classer_path, factors);
result_image = result_pro.draw_result(result_pro.process_result(det_result_array, proto_result_array), image.Clone());
if (!result_image.Empty())
{
pictureBox2.Image = new Bitmap(result_image.ToMemoryStream());
textBox1.Text = "推理耗时:" + (dt2 - dt1).TotalMilliseconds + "ms";
}
else
{
textBox1.Text = "无信息";
}
button2.Enabled = true;
}
private void Form1_Load(object sender, EventArgs e)
{
startupPath = System.Windows.Forms.Application.StartupPath;
model_path = startupPath + "\\yolo_edge_det.onnx";
classer_path = startupPath + "\\lable.txt";
// 创建输出会话,用于输出模型读取信息
options = new SessionOptions();
options.LogSeverityLevel = OrtLoggingLevel.ORT_LOGGING_LEVEL_INFO;
// 设置为CPU上运行
options.AppendExecutionProvider_CPU(0);
// 创建推理模型类,读取本地模型文件
onnx_session = new InferenceSession(model_path, options);//model_path 为onnx模型文件的路径
// 输入Tensor
input_tensor = new DenseTensor<float>(new[] { 1, 3, 800, 800 });
// 创建输入容器
input_container = new List<NamedOnnxValue>();
image_path = "1.jpg";
pictureBox1.Image = new Bitmap(image_path);
}
}
}
using Microsoft.ML.OnnxRuntime;
using Microsoft.ML.OnnxRuntime.Tensors;
using OpenCvSharp;
using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Linq;
using System.Security.Cryptography;
using System.Text;
using System.Web;
using System.Windows.Forms;
using static System.Net.Mime.MediaTypeNames;
namespace Onnx_Yolov8_Demo
{
public partial class Form1 : Form
{
public Form1()
{
InitializeComponent();
}
string fileFilter = "*.*|*.bmp;*.jpg;*.jpeg;*.tiff;*.tiff;*.png";
string image_path = "";
string startupPath;
string classer_path;
DateTime dt1 = DateTime.Now;
DateTime dt2 = DateTime.Now;
string model_path;
Mat image;
SegmentationResult result_pro;
Mat result_image;
SessionOptions options;
InferenceSession onnx_session;
Tensor<float> input_tensor;
List<NamedOnnxValue> input_container;
IDisposableReadOnlyCollection<DisposableNamedOnnxValue> result_infer;
DisposableNamedOnnxValue[] results_onnxvalue;
Tensor<float> result_tensors_det;
Tensor<float> result_tensors_proto;
private void button1_Click(object sender, EventArgs e)
{
OpenFileDialog ofd = new OpenFileDialog();
ofd.Filter = fileFilter;
if (ofd.ShowDialog() != DialogResult.OK) return;
pictureBox1.Image = null;
image_path = ofd.FileName;
pictureBox1.Image = new Bitmap(image_path);
textBox1.Text = "";
image = new Mat(image_path);
pictureBox2.Image = null;
}
private void button2_Click(object sender, EventArgs e)
{
if (image_path == "")
{
return;
}
button2.Enabled = false;
System.Windows.Forms.Application.DoEvents();
// 配置图片数据
image = new Mat(image_path);
int max_image_length = image.Cols > image.Rows ? image.Cols : image.Rows;
Mat max_image = Mat.Zeros(new OpenCvSharp.Size(max_image_length, max_image_length), MatType.CV_8UC3);
Rect roi = new Rect(0, 0, image.Cols, image.Rows);
image.CopyTo(new Mat(max_image, roi));
float[] det_result_array = new float[37 * 13125];
float[] proto_result_array = new float[32 * 200 * 200];
float[] factors = new float[4];
factors[0] = factors[1] = (float)(max_image_length / 800.0);
factors[2] = image.Rows;
factors[3] = image.Cols;
// 将图片转为RGB通道
Mat image_rgb = new Mat();
Cv2.CvtColor(max_image, image_rgb, ColorConversionCodes.BGR2RGB);
Mat resize_image = new Mat();
Cv2.Resize(image_rgb, resize_image, new OpenCvSharp.Size(800, 800));
// 输入Tensor
for (int y = 0; y < resize_image.Height; y++)
{
for (int x = 0; x < resize_image.Width; x++)
{
input_tensor[0, 0, y, x] = resize_image.At<Vec3b>(y, x)[0] / 255f;
input_tensor[0, 1, y, x] = resize_image.At<Vec3b>(y, x)[1] / 255f;
input_tensor[0, 2, y, x] = resize_image.At<Vec3b>(y, x)[2] / 255f;
}
}
//将 input_tensor 放入一个输入参数的容器,并指定名称
input_container.Add(NamedOnnxValue.CreateFromTensor("images", input_tensor));
dt1 = DateTime.Now;
//运行 Inference 并获取结果
result_infer = onnx_session.Run(input_container);
dt2 = DateTime.Now;
// 将输出结果转为DisposableNamedOnnxValue数组
results_onnxvalue = result_infer.ToArray();
// 读取第一个节点输出并转为Tensor数据
result_tensors_det = results_onnxvalue[0].AsTensor<float>();
result_tensors_proto = results_onnxvalue[1].AsTensor<float>();
det_result_array = result_tensors_det.ToArray();
proto_result_array = result_tensors_proto.ToArray();
resize_image.Dispose();
image_rgb.Dispose();
result_pro = new SegmentationResult(classer_path, factors);
result_image = result_pro.draw_result(result_pro.process_result(det_result_array, proto_result_array), image.Clone());
if (!result_image.Empty())
{
pictureBox2.Image = new Bitmap(result_image.ToMemoryStream());
textBox1.Text = "推理耗时:" + (dt2 - dt1).TotalMilliseconds + "ms";
}
else
{
textBox1.Text = "无信息";
}
button2.Enabled = true;
}
private void Form1_Load(object sender, EventArgs e)
{
startupPath = System.Windows.Forms.Application.StartupPath;
model_path = startupPath + "\\yolo_edge_det.onnx";
classer_path = startupPath + "\\lable.txt";
// 创建输出会话,用于输出模型读取信息
options = new SessionOptions();
options.LogSeverityLevel = OrtLoggingLevel.ORT_LOGGING_LEVEL_INFO;
// 设置为CPU上运行
options.AppendExecutionProvider_CPU(0);
// 创建推理模型类,读取本地模型文件
onnx_session = new InferenceSession(model_path, options);//model_path 为onnx模型文件的路径
// 输入Tensor
input_tensor = new DenseTensor<float>(new[] { 1, 3, 800, 800 });
// 创建输入容器
input_container = new List<NamedOnnxValue>();
image_path = "1.jpg";
pictureBox1.Image = new Bitmap(image_path);
}
}
}
参考
https://github.com/hpc203/TableDetection
百度网盘AI大赛-表格检测第2名方案 - 飞桨AI Studio星河社区