長久以來地下環境污染最不易即時察覺,往往需要經過一段長時間後,因需要土地整修才被人發現,往往污染程度已經十分嚴重,污染範圍也擴散滲透到難以探查。也因此在有探查的必要時,在人力、經費、工程上之浩大,也常隨污染擴大的程度而增加,因此有必要研究開發更簡便有效的科學方法來探查地下污染。目前非破壞性檢測最常運用的是透地雷達,但透地雷達檢測有其解析度的瓶頸和盲點,加上圖像的判讀分析解釋上需豐富經驗的專業人員來鑑別,所以更需要搭配其他更有效益的工具軟體才行。 本研究運用自行撰寫的MATLAB數位影像處理技術程式,處理透地雷達所得到原始實際案例污染影像資料,來進行判別地下有污染和無污染的區分,以證實MATLAB在影像處理技術上之可行性和準確度,藉以降低人為辨識上對透地雷達的盲點。
Underground pollution is unable to be detected in a timely manner, which is often detected after a long time during land reclamation while the pollution has already reached a serious degree, and the pollution range has widespread. Consequently, the input of labor, expenditure and engineering efforts is considerable in case of necessary detection, which increases along with the degree of pollution spreading. Therefore, it is necessary to study and develop a more efficient scientific method to detect underground pollution. Presently, the most commonly used non-destructive testing method is ground-penetrating radar detection, but there are some bottleneck and blind spots in terms of resolution in ground-penetrating radar detection. Moreover, interpretation, analysis and explanation of the images require experienced professionals. Therefore, more efficient application software is needed. This study used self-developed MATLAB digital image processing program to process the pollution image data of original and actual cases obtained through ground-penetrating radar detection, in order to determine the difference between pollution and non-pollution. The results verified the feasibility and accuracy of MATLAB application in image processing technology, and reducing the blind spots of ground-penetrating radar detection by artificial identification.