- 期刊
Determination of Optimum Frame Rates for Observation of Construction Operations from Time-Lapse Movies
並列摘要
Construction professionals have been using time-lapse movies in monitoring construction operations. However, some amount of detail is always lost in the interval between two consecutive frames in a time-lapse movie. This poses the question: By how much can the frame rate be lowered from the standard 30fps (frames per second) to allow for the accurate observation of construction operations from a time-lapse movie? This paper addresses the problem by establishing the optimum frame rates for observation of activities related to mortar mixing and block handling. The activities were first recorded at the standard rate of 30fps. Using the Adobe Premier Pro video editing software, the records were then segregated into still images from which 15 different time-lapse movies of various time intervals were generated. The movies were then shown to 25 Construction Managers. A structured questionnaire was employed to capture the level of accuracy with which Construction Managers could interpret the job site situation from each movie. The results suggest that 1fpm (frame per minute) is sufficient for the accurate tracking of labourers involved in mortar mixing while 1 frame in every 20 seconds is sufficient for accurate identification of number of cement bags used. However, for tracking number of blocks off-loaded, and those damaged, 1 frame in every 2 seconds is required.
延伸閱讀
- Wang, C. M. (2008). Multi-Frame Optical Flow Estimation and its Applications [doctoral dissertation, National Central University]. Airiti Library. https://www.airitilibrary.com/Article/Detail?DocID=U0031-0207200917355935
- Zeng, X., Chen, Z. Y., & Xiong, Z. (2009). Self-Recovery Algorithm Based on Motion Estimation of I-Frames for Compressed Video. 網際網路技術學刊, 10(5), 591-600. https://doi.org/10.6138/JIT.2009.10.5.14
- Wang, D. J., & Leou, J. J. (2001). A New Approach to Video Format Conversion Using Bidirectional Motion Estimation and Hybrid Error Concealment. Journal of Information Science and Engineering, 17(5), 763-777. https://doi.org/10.6688/JISE.2001.17.5.2
- Galayda, N. M. (2022). Full Duration Flare Detection, Classification and Region Extraction using Bidirectional Convolutional LSTMs [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU202204087
- 胡嘉錡（2009）。Rate-Distortion Optimization of H.264 SVC Internet Streaming via RTP with Receiver-driven Retransmission〔碩士論文，國立交通大學〕。華藝線上圖書館。https://doi.org/10.6842/NCTU.2009.01085