GGG-BenchmarkSfM: Dataset for Benchmarking Close-range SfM Software Performance under Varying Capturing Conditions
Description
The proposed dataset aims to benchmark the performance of SfM software under varying conditions - different environments, different lighting, image positions, camera setups, etc. Images of six objects are provided with varying shapes, sizes, surface textures and materials. The dataset is divided in two main parts, together with ReadMe files: - Objects and environments data - images from each of the objects both from indoor and outdoor environments are provided. - Capturing setups data - images from one of the objects are provided captured with different setups. Both with and without using a turntable, using one and multiple light sources and different amount of images All images are captured using Canon 6D DSLR camera. All images contain EXIF data with used camera parameters. A ground truth high resolution scanned of each of the objects is provided for verifying the accuracy of the SfM reconstructions.