Three Dimensional Reconstruction of Watermelon for Multimedia Traceability System

Abstract

Traceability system is an important technology for modern agriculture to improve the product quality and increase the brand influence. However, the current traceability system can only provide simply text data so that its effect is very limited. To overcome the shortcoming of text traceability system, we defined the new concept of Multimedia Traceability System (MTS). The multimedia traceability system uses computer technologies to collect, store, process, analyze, transfer and retrieve the data of text, image, graphic, video and voice, also establishes logical relationship among these data and integrates them into an interactive and multidimensional system. We proposed multimedia traceability framework and discussed the involved key technologies. Furthermore, we studied three dimensional reconstruction of fruit used in multimedia traceability system, taken watermelon as an example. We firstly took sequential pictures around watermelon from different angles, performed feature points matching using scale-invariant feature transform algorithm, then generated sparse point cloud and dense point cloud exploiting structure from motion method and multi-view stereo method, respectively. Consequently, texture mapping and model meshing were conducted by means of Poisson surface reconstruction approach. Finally, the accuracy and efficiency of reconstruction were analyzed by experiments.