The performance evaluation of multi-image 3D reconstruction software with different sensors

摘要

Three dimensional modelling of the objects is one the most discussed issues in Photogrammetry and Computer Vision. One of the factors affecting the accuracy of the obtained model in image-based methods is the software and algorithm to generate the model. Another important factor is the type of imaging sensor. Due to availability of cell phone sensors to the public, popularity of professional sensors and the advent of stereo sensors, a question raised about which imaging sensor can lead to more accurate and complete model. Although many research have been accomplished to identify a suitable software and algorithm to achieve an accurate and complete model, little attention has been paid to the type of imaging sensors and its effects on the quality of the final model. This paper aims to introduce an appropriate combination of a sensor and software to provide an accurate and complete model. To do this, different available software were compared and the most popular ones in each category were chosen. In this test, four small objects with distinct geometric properties were chosen and their accurate models as reliable true data were generated. Images were then captured using Fujifilm Real 3D stereo camera, Apple iPhone 5s and Nikon D3200 professional camera and three dimensional models of the objects were obtained using each of the software. Finally, a comprehensive comparison between the results showed that the best combination of software and sensors for generating three-dimensional models is directly related to the type of imaging sensors, the software and object shape. Generally better quantitative and qualitative results were obtained by using the Nikon D3200 professional camera.