Systemization of Visual Inspecting Processes for Finished Surfaces based on the Extracted Features of Them
Grant-in-Aid for Scientific Research (C)
|Research Institution||CHIBA UNIVERSITY|
HIGUCHI Shizuichi CHIBA UNIV., ENGRG., ASSOCIATE PROF., 工学部, 助教授 (50114319)
|Project Fiscal Year
1997 – 1998
Completed(Fiscal Year 1998)
|Budget Amount *help
¥2,200,000 (Direct Cost : ¥2,200,000)
Fiscal Year 1998 : ¥500,000 (Direct Cost : ¥500,000)
Fiscal Year 1997 : ¥1,700,000 (Direct Cost : ¥1,700,000)
|Keywords||Hairline finishing / Scratch / Image processing / 2 dimensional-discrete Fourier transform / Neighborhood operation / Visual inspecting process / 画像処理 / 2次元離散的フーリエ変換 / 近傍処理 / 特徴抽出|
It is useful to systemize the visual inspecting processes carried by skilled workers for finished surfaces. For that purpose, the hairline finished surfaces were inspected based on the extracted features of them. For the first step of estimating the hairline finished surfaces, the scratch which had been left on the hairline pattern owing to inadequate finishing, was first detected. The difficulty of detecting a scratch in the hairline finishing is in separating the scratch from the hairline pattern. The hairline images for inspecting were taken by a CCD camera and analyzed by a personal computer. 2 dimensional discrete - Fourier transform(2D - DFT)was applied for estimating the hairline images obtained by the CCD camera. As a result, it was found that there appears characterized spatial frequency distribution in the spacing frequency area, when their was a scratch in the hairline. In order to estimate characteristics of 2D - DFT results, a fan- shaped filter was utilized. The amount of
the power which was included in the fan- shaped filter was calculated consequently and it was found that the fluctuations of the amount of power included in the fan- shaped filter were related to the size and direction of the scratch in the hairline finishing.
Moreover, the position and size of scratch in the hairline image were recognized. For that purpose, two kinds of operators were considered from the mean distance of hairline pattern and the directions of scratch. It was confirmed that the size and the position of scratch could be clearly recognized by the neighbor hood operation of the hairline image using them. By applying this method in various cases, the effectiveness of this method was confirmed.
Research Output (3results)