在本論文中,我們提出一個基於影像處理結合移動平均的方案,建構在火災辨識上。在攝影機所取得的影像進行影像處理後,並加入移動平均,用於偵測場景中是否有火焰。 本論文分為三個部分,第一部分,我們先利用攝影機所取得的影片截取成每秒間隔的單張圖像,之後我們將截取完的圖像使用軟體OpenCV來獲得圖像的灰階變化,為了分析更加準確,將移動平均加入灰階像素值中。第二部分,也是先利用攝影機所取得的影片截取成每秒間隔的單張圖像,將截取完的圖像使用軟體OpenCV來獲得圖像的RGB色彩空間變化,再將RGB像素值中加入移動平均。最後的部分,是實驗模擬結果,我們可以看出灰階和RGB加入移動平均後,當我們在不同的地點,可以偵測出是否有火焰在場景中,皆能達到我們要的結果。 本文的研究貢獻如下: 安全性: 提供給一般住戶使用,利用攝影機進行即時的偵測,可以火勢變大之前,降低給消防人員救援的危險性,以及減少火災帶來的傷害和死亡率。 低成本: 利用現有的攝影機就能進行偵測,不需要額外加裝,有效節省成本。 趨勢性: 隨著現今科技的進步,將攝影機和火災偵測做結合,達到智慧安全。
In this thesis, we propose a scheme based on image process incorporation with moving averages to construct a model for fire identification. After we processed the video which was obtained by the camera, a moving average is added to detect whether there are flames in the scene. This thesis is divided into three parts. In the first part, we first use the video obtained by the camera to intercept a single per second interval images. Once we finish intercepting images, we will use the software OpenCV to obtain the grayscale variation. In order to make the analysis more accurate, we add the moving average to the grayscale variation values. In the second part, also use the video obtained by the camera to intercept a single per second interval images, we will use the software OpenCV to obtain the RGB color space variation, then add the moving average to the RGB variation values. The final part is the experimental simulation results. We can see that when grayscale and RGB are added to the moving average, and when we are in different scenes, we can detect whether there is a flame in the scene, and we can achieve the results we want. In this thesis, we have some contributions as follows: Security: This thesis can be provided for general household. The camera with real-time detection can reduce the risk of rescue for firefighters before the fire becomes serious as well as the damage and mortality caused by fire. Low cost: This thesis can use the existing camera to detect, no additional installation is required, which effectively saves costs. Trending: This thesis combines cameras and fire detection to achieve smart safety.