Magnetic tile as a kind of product of mass production and wide application in electronic motors, and defects detection is a major issue in its production line. In this paper, a machine vision method based on black-stripe projection is presented to deal with this issue. Because of magnetic tile surface with black colors, rough structure and complex grinding textures, we abandon intensity imaging and resort to light sectioning methods which provides more reliable and abundant surface information. In order to suppress the speckle diffraction effect caused by laser light source, we used the light-emitting diode (LED) with incoherent characteristics. The black-stripe images were captured by a high-speed camera. A fast algorithm was developed to extract and compare both edges of the black-stripe, which could detect defects and eliminate the effects of vibrations. The experimental results show that the simple and fast processing method proposed in this paper can detect the structural defects such as micro pits and micro cracks.
A visual automatic detection method is proposed for defect detection on end surface of ferrite magnetic tile to tackle the disadvantages generated by human work which has low efficiency and unstable accuracy. Because the defects on end surface of ferrite magnetic tile with dark colors and low contrasts are negative for defect detection, uniform illumination is provided by LED light source and a dedicated optical system is designed to extract defects conveniently. The approach uses comparison of the fitting and actual edge curves to detect defects mainly with most defects located on the edge. Firstly improved adaptive median filter is used as the image preprocessing. Subsequently the appropriate threshold is calculated by Otsu algorithm based on the extreme points in the gray-level histogram to segment the preprocessing image. Then the Sobel operator can be used to extract the edge of end surface precisely. Finally through comparing the ideal fitting and actual edge curves of end surface, to detect the defects with some relevant features. Experimental results show that the proposed scheme could detect defects on the end surface of ferrite magnetic tile efficiency and accurately with 93.33% accuracy rate, 2.30% false acceptance rate and 8.45% correct rejection rate.
Advanced image sensor and powerful parallel data acquisition chip can be used to collect more detailed and comprehensive light field information. Using multiple single aperture and high resolution sensor record light field data, and processing the light field data real time, we can obtain wide field-of-view (FOV) and high resolution image. Wide FOV and high-resolution imaging has promising application in areas of navigation, surveillance and robotics. Qualityenhanced 3D rending, very high resolution depth map estimation, high dynamic-range and other applications we can obtained when we post-process these large light field data. The FOV and resolution are contradictions in traditional single aperture optic imaging system, and can’t be solved very well. We have designed a multi-camera light field data acquisition system, and optimized each sensor’s spatial location and relations. It can be used to wide FOV and high resolution real-time image. Using 5 megapixel CMOS sensors, and field programmable Gate Array (FPGA) acquisition light field data, paralleled processing and transmission to PC. A common clock signal is distributed to all of the cameras, and the precision of synchronization each camera achieved 40ns. Using 9 CMOSs build an initial system and obtained high resolution 360°×60° FOV image. It is intended to be flexible, modular and scalable, with much visibility and control over the cameras. In the system we used high speed dedicated camera interface CameraLink for system data transfer. The detail of the hardware architecture, its internal blocks, the algorithms, and the device calibration procedure are presented, along with imaging results.
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.