Stereo Vision is the key to 3D reconstruction, autonomous navigation, and non-contact ranging etc. semi-global stereo matching (SGM) is one of the stereo matching algorithms to obtain dense depth map, which is with both relatively high accuracy and efficiency. This paper describes a double feature fusion SGM algorithm, which is a fusion of the mean absolute difference (MAD) and Census transform as intensity to make cost calculation. A combination of MAD for the sensitivity to the texture and Census transform for the advantages of adaptive illumination change. This method is especially suitable for parallel computing processing platform and can be used in practical scenarios which have a higher demand of real-time performance.
Infrared image quality is a key factor of object detection. Stripe nonuniformity is very typical in the staring infrared focal plane array (IRFPA). In this paper, we propose a novel high-space-frequency correction method to eliminate the stripe nonuniformity. The kernel ideal is to eliminate the high-space-frequency part of stripe nonuniformity and retain its low-space-frequency part which can reduce unwanted ghosting artifacts. Firstly, the spatial characteristic of stripe nonuniformity is discussed, then correction parameters are computed based on the spatial high frequency part of image. Experimental results show the proposed mehtod can compute adaptive correction parameters of each readout channel and obtain a reliable stripe nonuniformity reduction.
For infrared focal plane array sensors, imagery is degraded by a number of phenomena during signal acquisition, particularly including under-sampling and detector non-uniformity. In this paper, we propose an efficient framework which combines neural network non-uniformity correction with image registration for removing structured and non-structured noise and increasing spatial resolution. To achieve this, we sequentially improve the image quality in two steps: primarily, removing the structured and non-structured noise based on neural network theory, and achieving registration using an iterative gradient-based registration technique. Experimental results are presented to demonstrate the effectiveness of the proposed algorithm. By using our method, the shifts between acquired frames are estimated precisely and the quality of reconstructed image is improved.
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.