Cracks on sheet metals can significantly affect the overall strength. Crack detection during manufacturing is, thus, an important process for the quality assessment on a press line. Deep learning, a data-driven structure, has been extensively used to detect cracks on various surfaces. In this study, a crack detection technique for a press line using Retina Net and a novel data augmentation method is proposed, which mainly focuses on three steps, shape acquisition, style transfer, and edge fusion. First, the shapes of crack on different materials are extracted. Then, images are created by providing metal crack textures to those shapes using a fusion network with a relatively small number of real crack images. Real crack images are captured from a sheet metal forming line. Training data can be enriched using the proposed data augmentation method. Validation experiments are conducted to demonstrate the effectiveness of the proposed crack detection and data augmentation techniques.
The computer vision-based measurements offer the superior capabilities over traditional sensing systems, including high spatial resolutions, no mass-loading effects, and low cost. This capability allows the vision-based technique to be widely applied to the damage detection practice as a more efficient way compared to conventional methods. The recently developed phase-based motion processing technique can measure the displacement signals with high accuracy and noise robustness. In this study, an automated damage identification and localization technique based on the phase-based motion processing is proposed. The local phase of the object is extracted using a optimal steerable filter. The modal parameters are then obtained after performing the morphological operation on the edge image in conjunction with the phase signals. Damage is finally localized and assessed by the features extracted by the identified modal parameters. Several experiments are carried out to validate the proposed technique on a 5-story structure under different bolt losing conditions. The experimental results show that the proposed technique can automatically detect and evaluate the damage with high accuracy.
The advanced displacement measurement technique using computer vision has shown several advantages, such as high spatial resolution and no mass loading effect, compared to conventional sensing techniques. However, the accuracy and robustness of vision-based techniques are subjected to the various conditions, including uneven illumination and insufficient lighting. This study introduces an accurate 2-dimension displacement measuring technique with high robustness to the illumination change, which uses two complex Gabor filters and a specially designed marker. The linear phase can be generated around the marker by optimizing the filter parameter for accurate motion estimation. The nonlinearity caused by the complex conditions, such as low light and uneven illumination, can also be reduced by emphasizing marker features. Phase-based optical flow is further employed to extract the displacement based on the extracted phase. The measurement performance is compared with Laser Doppler Velocimetry (LDV) to validate the proposed technique under various lighting conditions and its robustness is demonstrated. The proposed technique is also applied to different structures to show the ability of measuring high-accuracy displacement signals under various conditions.
Strabismus is an eye movement disorder that the eyes do not properly align with each other when looking at an object. This disorder is usually caused by muscle malfunctions, nerve problems or injuries. Currently, the ophthalmic prism with two nonparallel planes is used to diagnose the strabismus angle. The light into one eye is refracted when passing through the prism, which adjusts both eyes to looking forward. The strabismus angle is then identified after checking the parameter of the prism. However, the whole process is operated depending on the doctors’ experience which shows somewhat low efficiency and low accuracy. In this study, an automated strabismus diagnosis technique using VR device is developed. A specially-designed VR is built to simulate the normal strabismus diagnosis steps, in which screens are controlled to change alternately between on and off. The eye motions are tracked by two IR cameras by an image-processing based pupil tracking technique. After tracking the motion of the pupil, the position information is converted to the strabismus angle by considering the eyeball diameter. With this process, the strabismus angle is accurately and automatically identified using a unique feature recognition technique. To demonstrate the performance of this technique, experiments are carried out on various persons, including strabismus patients. The results are compared to the doctor’s diagnosis. The results show that this technique could identify the strabismus angle with high accuracy and high efficiency.
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.