基於自行車正面騎乘視訊之姿勢分析及其在常態訓練之應用

本論文提出的兩項技術，針對與智慧型裝置整合的自行車體能訓練應用軟體所遇到的問題，提出具體的改善方法。在自行車上，踏頻數據與騎乘者的體能消耗程度有絕對的關係，是運動員或者單車愛好騎士所關心的重要騎乘參數，但用於智慧型裝置中計算踏頻數據的應用軟體，卻需要加裝額外的感測器才能估算踏頻，尚有移植性低、硬體成本高及裝置之間訊號傳輸穩定性不足的問題；故我們開發出「基於視訊的踏頻量測方法」，希望僅以智慧型裝置內建的攝相機及微處理器完成踏頻數據的取樣、計算及結果呈現，改善傳統踏頻計額外硬體成本的問題；實驗結果顯示此方法量測出的踏頻誤差相當小，在未來可以取代智慧型裝置中踏頻計算軟體。我們也根據得到的踏頻數據，研擬一套新的運動訓練方案，可以根據騎乘者當下的生理狀態及預先設定的訓練目標，在運動過程中即時地給予適當的建議，修正騎乘者的運動強度或運動時間，以避免過度運動造成身體傷害或運動不足而沒有達到預期訓練目標的問題。與傳統用於智慧型裝置中，基於心跳訊號制定的訓練方案相比，我提出以HRV及踏頻的訓練方案制訂方法，可以改善智慧型裝置的訓練方案準確度過低的問題，以提升其參考價值。

關鍵字

視訊處理；自行車踏頻計；訓練方案

並列摘要

We propose two techniques to solve the cycling training apps problems on smart devices. Cycling cadence is the most important parameter which directly reflects cyclist’s physical power consumption and exercise intensity. To estimate the cycling cadence parameter, we can only use induction device as a way to measure the cadence data; Besides the high cost problem, it is also unstable in wireless connection between devices and low portability. Thus, we propose vision-based cadence measurement method to estimate cadence data by ruling cyclist’s body dynamics. The accuracy cadence results show that the method we proposed can replace the traditional cadence meter in the future. Based on the cadence data we measured, we develop a new training program, which use heart rate and cadence data as system’s input and cycling suggestion like recommended exercise intensity or training time as system’s output. These feedbacks to cyclist in order to make ensure cyclist can achieve their training purpose and protect them away from over exercise. Compared with the traditional heart-rate-based training program in apps, the method pedaling and HRV based training program we proposed can promote training program’s consultant value on smart devices.

並列關鍵字

Training program ； Vision processing ； Cadence meter

參考文獻

[1] H. Tanaka, K. D. Monahan, and D. R. Seals, “Age-predicted maximal heart rate revisited,” JACC, 2001, pp. 153–156.

[2] H. Mitchell, et al., “ACSM’s guidelines for exercise testing and prescription,” 7th ed., Ed. Philadelphia: Lippincott Williams & Wilkins, 2006.

[4] J. Canny, “Computational approach to edge detection,” PAMI, 1986, pp. 679–698.

[7] R. Duda and P. Hart, “Use of the hough transformation to detect lines and curves in pictures,” Commun. ACM, 1972, pp. 11-15

[8] P. Viola and M. Jones, “Robust real-time object detection,” in Proc. ICCV, 2001.

國際替代計量

基於自行車正面騎乘視訊之姿勢分析及其在常態訓練之應用

未授權

主題瀏覽