Omnidirectional camera calibration in python

Last update: Nov 22, 2022

Related tags

Deep Learning py-omnicalib

Overview

Omnidirectional Camera Calibration

Key features

pure python
initial solution based on A Toolbox for Easily Calibrating Omnidirectional Cameras (Davide Scaramuzza, Agostino Martinelli, Roland Siegwart)
Optimized solution via Levenberg–Marquardt algorithm on reprojection error
Lie algebra based extrinsics optimization

(undistort example)

Usage

Corner detection (see python -m omnicalib.detect --help for detailed argument description)

python -m omnical.detect --chessboard <rows> <columns> <square-size> --max-dim <max-dim> --threads <threads> <image-folder>

The file detections.pickle is written, which contains a pickled dictionary with the following format

{
    'detections': {
        <image_path_0>: {'image_points': <image_points>, 'object_points': <object_points>},
        ...
}

Image paths (pathlib.Path) are absolute
image_points are torch.Tensor with shape N x 2 and dtype torch.float64
object_points are torch.Tensor with shape N x 3 and dtype torch.float64 and third colum, the z coordinate, all zeros

If an external corner detection method is used, this file can simply be written manually.

Calibration (see python -m omnicalib --help for detailed argument description)

python -m omnical --degree <degree> detections.pickle

Result

A calibration.yml file (see example_calibration.yml) containing

extrinsics as list of 3 x 4 matrices
poly_incident_angle_to_radius, a polynom that converts incident angle to image radius
poly_radius_to_z, a polynom that converts radius to z component of view vector
principal_point in pixel

Method

For a detailed description see method.md.

Simulation

Play with the calibration on simulated data with the provided jupyter lab notebook. It generates randomized data for the following ideal projection models

equidistant
stereographic
orthographic
equisolid

, see Large Area 3D Human Pose Detection Via Stereo Reconstruction in Panoramic Cameras for a detailed description.

Example

Jupyter lab notebook, example results

You might also like...

Improving Calibration for Long-Tailed Recognition (CVPR2021)

19 Apr 28, 2021

Pytorch Implementation of Spiking Neural Networks Calibration, ICML 2021

SNN_Calibration Pytorch Implementation of Spiking Neural Networks Calibration, ICML 2021 Feature Comparison of SNN calibration: Features SNN Direct Tr

60 Dec 27, 2022

S2-BNN: Bridging the Gap Between Self-Supervised Real and 1-bit Neural Networks via Guided Distribution Calibration (CVPR 2021)

S2-BNN (Self-supervised Binary Neural Networks Using Distillation Loss) This is the official pytorch implementation of our paper: "S2-BNN: Bridging th

52 Dec 24, 2022

[CVPR 2022] Official code for the paper: "A Stitch in Time Saves Nine: A Train-Time Regularizing Loss for Improved Neural Network Calibration"

MDCA Calibration This is the official PyTorch implementation for the paper: "A Stitch in Time Saves Nine: A Train-Time Regularizing Loss for Improved

21 Dec 22, 2022

Omnidirectional camera calibration in python

Related tags

Overview

Omnidirectional Camera Calibration

Key features

Usage

Result

Method

Simulation

Example

You might also like...

Improving Calibration for Long-Tailed Recognition (CVPR2021)

Pytorch Implementation of Spiking Neural Networks Calibration, ICML 2021

S2-BNN: Bridging the Gap Between Self-Supervised Real and 1-bit Neural Networks via Guided Distribution Calibration (CVPR 2021)

the code for paper "Energy-Based Open-World Uncertainty Modeling for Confidence Calibration"

We utilize deep reinforcement learning to obtain favorable trajectories for visual-inertial system calibration.

Improving Calibration for Long-Tailed Recognition (CVPR2021)

Source code of NeurIPS 2021 Paper ''Be Confident! Towards Trustworthy Graph Neural Networks via Confidence Calibration''

CSAC - Collaborative Semantic Aggregation and Calibration for Separated Domain Generalization

[CVPR 2022] Official code for the paper: "A Stitch in Time Saves Nine: A Train-Time Regularizing Loss for Improved Neural Network Calibration"

Releases(1.0.1)

1.0.1(Mar 1, 2022)

1.0.0(Dec 21, 2021)

Owner

Thomas Pönitz

Source code for CIKM 2021 paper for Relation-aware Heterogeneous Graph for User Profiling

Reproducing code of hair style replacement method from Barbershorp.

Text to image synthesis using thought vectors

Code and data form the paper BERT Got a Date: Introducing Transformers to Temporal Tagging

TransPrompt - Towards an Automatic Transferable Prompting Framework for Few-shot Text Classification

Code of paper: "DropAttack: A Masked Weight Adversarial Training Method to Improve Generalization of Neural Networks"

Code for the paper "Adversarial Generator-Encoder Networks"

PyTorch implementation of PSPNet

GEA - Code for Guided Evolution for Neural Architecture Search

A curated list of long-tailed recognition resources.

Python 3 module to print out long strings of text with intervals of time inbetween

This repository contains pre-trained models and some evaluation code for our paper Towards Unsupervised Dense Information Retrieval with Contrastive Learning

Graph Robustness Benchmark: A scalable, unified, modular, and reproducible benchmark for evaluating the adversarial robustness of Graph Machine Learning.

Implementation of PyTorch-based multi-task pre-trained models

Contrastive Feature Loss for Image Prediction

Repositório criado para abrigar os notebooks com a listas de exercícios propostos pelo professor Gustavo Guanabara do canal Curso em Vídeo do YouTube durante o Curso de Python 3

Efficient Speech Processing Tookit for Automatic Speaker Recognition

DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time

Fast Neural Representations for Direct Volume Rendering

Deep Semisupervised Multiview Learning With Increasing Views (IEEE TCYB 2021, PyTorch Code)