固定式起重機(Cranes)為我國法定之危險性機械(Dangerous Machine)，天車是固定式起重機之一種，指的是在室內進行垂吊上昇、放下、運輸重物的操作機具，其運輸動作和動線是由升降捲動馬達、驅動馬達及軌道所組成，其操作動線和吊重都有所限制，如果操作不慎所衍生之工安事件，輕則造成作業人員受傷，重則導致財物損失及作業人員之重大傷亡。因此在操作時必須非常謹慎小心，對於荷重大小、路徑及作業四周環境等，均要確實瞭解掌握，才能確保作業安全與避免災害發生。 本研究是針對天車系統（Overhead Crane System）設計預防碰撞的安全警報裝置，利用視覺監控判別吊掛物體與人員（操作者、現場工作者）之距離關係，所採用影像處理技術包含移動物體辨識、邊緣檢測、影像特徵比對等，同時進一步利用雷射距離感測器，預防吊掛物體可能撞擊人員時，即時發送警報以提醒操作者、及現場工作者，必要時也能強制切斷動力源，達到安全預防效果。 The stationary crane is defined as a dangerous machine in the law of Labor Safety and Health in R.O.C. The overhead crane is a kind of stationary cranes. It is an operating machinery to hang up/or down and transport the heavy goods in the industry. The overhead crane is composed of winding motor, driving motor, track and a control panel. It will be easy to get hurt if the operator makes a mistake in the regular operation. Therefore, the operator must be very careful to handle all the lifting processes in order to ensure safety and avoid any disasters. The objective of this research is to design a safety monitoring system for an overhead crane in order to prevent any collision between the moving hanged goods and persons. The distances between persons and goods are detected by a machine vision system. The image techniques are object feature recognition, object edge detection and image feature maching. Two distance scanning sensors are installed in the monitory system to confirm all the detecting moving object in order to avoid any disasters. If there will be any emergency dangerous, the monitory system will send an alarm to remind the operator or workers simultaneously. Finally, the power of the crane system will be cut off in order to prevent any collision.
