Binocular stereo vision technology plays an essential role in the intelligent manufacturing system due to the advantages of high accuracy and non-contact. However, for the measurement of components with highly reflective surfaces such as metal and ceramic, the specular reflections affected by the complex light field lead to the failure of feature matching and the decrease of measurement accuracy. This paper proposes an imaging strategy for binocular vision and a high dynamic image processing method to suppress the effect of specular reflection for stereo matching. Firstly, the mechanism of highlight generation in the image is analyzed by combining the illumination reflection theory of the BRDF model. Then, a binocular vision system with a parallel optical axis is built to capture images under different illumination conditions. The image processing algorithm of high dynamic range image fusion is studied, and an algorithm based on tone mapping and weight fusion is implemented to remove the high-lights. Finally, an experiment was performed to verify the effectiveness of the proposed method by a robust and fast matching algorithm. An image evaluation method based on BRISQUE further demonstrates the effectiveness of the method. Compared to the original images, the quality score of the HDR images is lower, which means that the processed images are of better quality. Moreover, the method provides an increase of 23.33% in image matching accuracy, which verifies the availability applied in the measurement of highly reflective surface components.
KEYWORDS: Clouds, Denoising, Digital filtering, Signal to noise ratio, Image segmentation, Optical filters, Visualization, Data acquisition, Principal component analysis
The assembly gap between components is very vital for the evaluation of assembly quality of aircrafts. Due to the limits of gap size and operation space, the assembly gap needs to be indirectly calculated by the measurements of surface of components instead of plug gauge test. However, the surface constituted of point cloud is usually mixed with different types of noise ,which severely affects the evaluation of assembly gap. To remove these different types of noise simultaneously with high efficiency, a classified denoising method combining with an improved bilateral filtering and median filtering was proposed. Firstly, based on the principal component analysis, a new coordinate system was established to achieve the homogeneity of coordinates of point cloud. Then, an improved median filtering method on the basis of region segmentation (RSMF) was used to remove large-scale noise. Next, the fast bilateral filtering method based on threshold segmentation (TSBF) was proposed to remove small-scale noise. Finally, a measurement experiment of aircraft component was performed to verify the effectiveness of the proposed method. Experimental results showed that the proposed method could not only reduce measurement error including RMSE (Root Mean Square Error), but also improve SNR (Signal Noise Ratio) and PSNR (Peak Signal to Noise Ratio) of point cloud data.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.