การวัดสัดส่วนร่างกายสำหรับแนวทางการเลือกขนาดจักรยานด้วยการประมวลผลภาพถ่าย

Abstract

Currently, the road bike used popular. The design, coupled with low or dropped handlebars, requires the rider to bend forward more that will have diaseas, such as back pain. The important thing other than position while riding is Road bike frame. This paper proposed a method of bicycle size computation frome rider image. This method consist of 7 steps: Prepare subject, Image acquisition, Body segmentation, Body part detection, bike geometry computation, bike geometry computation with condition and fitting bike for sized between rider and bicycle. Proformance test have 5 step. The experimental result, the result will be compared with result by bicycle expert selection. The precision of human body part of propose method is 98.96%. The precision of tube length of road bike is 98.73%. The precision of bike frame selection is 51.22%. The precision of bike frame selection with smaller condition is 51.22% and The precision of equipment to fitting bicycle : stem is 97.57%, Crankarm is 98.44% and Saddle 96.92% The advantage of this method are only one image in needed to compute the proper bicycle, do not waste the time and bike fitting processing can be achieved easier.