HSV色彩模型應用於手持裝置鋼材鏽蝕影像辨識系統開發

Abstract

橋梁工程相關之研究在我國營建產業的地位舉足輕重，我國目前對「鋼橋梁」之鏽蝕程度是否達到需要重漆之標準，為百分比上之定義。然而百分比為量化數據，需要經過專業的方法衡量，否則容易受到主觀的影響左右。 故本研究結合影像辨識的概念，目標為設計一能夠讀取照片並自動判斷鏽蝕、計算其比例之系統，並將之與智慧型可攜裝置結合，產出手機應用程式系統。系統原理為透過 HSV 色彩模型，定義「鏽蝕」之區間範圍並進行判斷辨識。 在研究最後，產出系統將比較前人利用邊緣檢測之演算法所建立之系統，發現在辨識效率及運算速度上，均獲得良好之成效。

The research related to bridge engineering plays an important role in Taiwan's construction industry. In Taiwan, the percentage of the degree of rust is considered as the standard of repainting steel bridge. Howerer, percentage is quantitative data, which needs to be measured by professional methods instead of subjective judgment. Therefore, this study combines the concept of image recognition with the goal of designing a system, which can read photos and automatically determine the rust and calculate the proportion. After that, this research combines it with a smart portable device to produce a mobile application system. The system principle is to define the range of 'corrosion' through the HSV color model and to judge and identify. At the end of this research, the output system will be compared to the system which uses edge detection algorithm, and find that good results are obtained in terms of identification efficiency and computation speed.