3D Reconstruction of Fruit Shape based on Vision and Edge Sections

3D Reconstruction of Fruit Shape based on Vision and Edge Sections


DOI: 

https://doi.org/10.30564/jeisr.v4i1.4585

Abstract

The fruit industry has been known as one of the largest businesses in Malaysia, where most of the fruits pass through the peeling process well in advance before the final product as juice in a bottle or slices in a can. The current industrial fruit peeling techniques are passive and inefficient by cutting parts of the pulp of the fruit with peels leading to losses. To avoid this issue, a multi-axis CNC fruit peeler can be used to precisely peel the outer layer with the guidance of a 3D virtual model of fruit. In this work, a new cost-effective method of 3D image reconstruction was developed to convert 36 fruit images captured by a normal RGB camera to a 3D model by capturing a single image every 10 degrees of fruit rotation along a fixed axis. The point cloud data extracted with edge detection were passed to Blender 3D software for meshing in different approaches. The vertical link frame meshing method developed in this research proved a qualitative similarity between the output result and the scanned fruit in a processing time of less than 50 seconds.

Keywords: 

3D Reconstruction, Machine vision, Fruit processing

Comments

Popular posts from this blog

𝐉𝐨𝐮𝐫𝐧𝐚𝐥 𝐨𝐟 𝐀𝐭𝐦𝐨𝐬𝐩𝐡𝐞𝐫𝐢𝐜 𝐒𝐜𝐢𝐞𝐧𝐜𝐞 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 | 𝐕𝐨𝐥𝐮𝐦𝐞 𝟎𝟔 | 𝐈𝐬𝐬𝐮𝐞 𝟎𝟑 | 𝐉𝐮𝐥𝐲 𝟐𝟎𝟐𝟑

𝗝𝗼𝘂𝗿𝗻𝗮𝗹 𝗼𝗳 𝗘𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝗮𝗹 & 𝗘𝗮𝗿𝘁𝗵 𝗦𝗰𝗶𝗲𝗻𝗰𝗲𝘀 | 𝗩𝗼𝗹𝘂𝗺𝗲 𝟬𝟰 | 𝗜𝘀𝘀𝘂𝗲 𝟬𝟮 | 𝗢𝗰𝘁𝗼𝗯𝗲𝗿 𝟮𝟬𝟮𝟮

𝐉𝐨𝐮𝐫𝐧𝐚𝐥 𝐨𝐟 𝐀𝐭𝐦𝐨𝐬𝐩𝐡𝐞𝐫𝐢𝐜 𝐒𝐜𝐢𝐞𝐧𝐜𝐞 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 | 𝐕𝐨𝐥𝐮𝐦𝐞 𝟎𝟔 | 𝐈𝐬𝐬𝐮𝐞 𝟎𝟐 | 𝐀𝐩𝐫𝐢𝐥 𝟐𝟎𝟐𝟑