以深度影像為基礎之健身器材輔助系統設計

Abstract

隨著社會的發展，越來越多人使用健身器材當成主要的運動方式。而健身器材的使用過程中，常需要健身教練的幫助來正確的使用器材。然而，由坊間的健身中心的發展來看，使用者的成長速度卻快於健身教練的增加速度。有鑑於此，本論文以比較不易受環境因素影響的深度影像為基礎，建立一個健身器材的即時輔助系統。希望透過一即時的使用者畫面回饋系統，讓健身器材的使用者，能在使用健身器材的同時，同步的接收到本論文輔助系統的姿勢建議。經過本論文的實際實驗測試後，本輔助系統不僅可以讓健身器材的使用者預防受傷，也更進而達到增進使用健身器材的效率。 本論文的主要貢獻：第一，在本論文建立一個非接觸式裝置之深度影像為基礎的即時視覺輔助系統。改善過去因環境光線的影響資料影像的缺點。第二，此系統被創新的運用於健身器材的輔助之上，並且利用獲取使用者空間中三維的位置資訊，進而改善健身教練過去因為人眼視角的限制而忽略的姿勢糾正錯誤。

More and more people use the fitness equipment to replace of regular exercises. And the equipment users often need the help of the fitness trainers to modify their pose. However, the pace of the gain in the number of fitness trainers is much slower than that in the number of users. In this paper, we build a real-time fitness equipment assistant system which is based on the depth image. We hope the equipment users can receive some real-time post modification feedback through this instant feedback system while they are using the equipment. And that is designed not only to prevent the sport injuries, but to reach the higher efficiency of the equipment usage. The following are contributions of this paper. First, we use a non-contact measurement equipment to get the depth image data and that for building a real-time feedback system and that improve the weakness of traditional color image. Second, this system is applied to the assistant system of the fitness equipment innovatively. This system is also designed to improve the mistakes which are caused by the fitness trainer's carelessness resulted from the limit of human eyes.