Fruit shape estimation from stereoscopic images in real time

本研究旨在改进水果自动检测和分类的过程。 现有的立体视觉算法提取深度信息是不适合真实的时间计算。 随着具有高度并行性和局部性的算法的发展，现场可编程门阵列的复杂性增加和成本降低沿着， 执行真实的所需计算的可能性。 这个项目的目的是调查适用于真实的时间硬件实现与应用到水果形状估计它真实的时间在文献中的各种立体视觉算法。

The research aims at improving the process of automatic fruit inspection and classification. Existing stereo vision algorithms to extract depth information are unsuitable for real time calculations. The increasing complexity and reducing cost of field programmable gate arrays along with the development of algorithms that have a high degree of parallelism and locality has created the possibility of performing the calculations required in real time. This projects aims to investigate the suitability of the various stereo vision algorithms available in the literature for real time hardware implementation with application to fruit shape estimation it real time.