# surface_match

**Repository Path**: Little_rock/surface_match

## Basic Information

- **Project Name**: surface_match
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: BSD-2-Clause
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-02-11
- **Last Updated**: 2025-02-11

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Surface Match: fast and robust point pair feature

## highlights:
- score based multi-instance
- OpenMP & SIMD based concurrency
- build kdtree only once
- pose refined with icp
- support model save/load

## usage:
```C++
#include <ppf.h>
int main(int argc, char *argv[]) {
    ppf::PointCloud model;
    ppf::readPLY(argv[ 1 ], model);
    ppf::PointCloud scene;
    ppf::readPLY(argv[ 2 ], scene);
    scene.viewPoint = {620, 100, 500};

    {
        ppf::Detector detector;
        detector.trainModel(model, 0.04f);
        detector.save("1.model");
    }

    {
        std::vector<Eigen::Matrix4f> pose;
        std::vector<float>           score;
        ppf::Detector detector;
        detector.load("1.model");
        detector.matchScene(scene, pose, score, 0.04f, 0.1f, 0.2f);
        
        for (int i = 0; i < pose.size(); i++) {
            std::cout << pose[ i ] << std::endl;
            std::cout << score[ i ] << std::endl;
        }
    }
    return 0;
}
```
## sample:
![gear](img/gear.png)

download [ gears model](https://download.ensenso.com/s/sampledata/download?files=gear.stl) (convert to ply first) and 
[gears scene](https://download.ensenso.com/s/sampledata/download?files=gear_n35.ply)