Development of the largest inscribed rectangle detection system for irregular shaped stones based on machine vision method

碩士 === 國立臺北科技大學 === 製造科技研究所 === 107 === In the manufacturing industry, there are certain issues associated with cutting raw materials. To obtain the largest rectangular area of a stone, the stone is cut by employing conventional visual methods based on measurements. However, such methods reduce the...

Full description

Bibliographic Details
Main Authors: LUO, YU-TING, 羅宇廷
Other Authors: YEH, SYH-SHIUH
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/29chq7
Description
Summary:碩士 === 國立臺北科技大學 === 製造科技研究所 === 107 === In the manufacturing industry, there are certain issues associated with cutting raw materials. To obtain the largest rectangular area of a stone, the stone is cut by employing conventional visual methods based on measurements. However, such methods reduce the availability and efficiency of the stone. Hence, a rotation histogram method with time complexity of is proposed as a detection technique for capturing images of irregularly shaped stones such as granite, marble, shale, and voided stone contours. The acquired images were rotated counterclockwise from 0° to 89° for determining the optimal angle to obtain the largest inscribed rectangle in a stone. Additionally, the four vertex coordinates of the largest inscribed rectangle were identified by applying the pointPolygonTest function. All results are found to reside within the contour, and they improve the calculation accuracy of the rotation histogram method. Furthermore, it is observed that the rotation histogram method is superior to the traversal center diffusion method because its utilization rate is higher by 3.27%, and its computation time is reduced by 64.83%.