Search

There is a need for the effective weld seam profile extraction method to realize the automatic and intelligent welding process with thick steel plates based on laser vision sensing. In this paper, a method was proposed to identify the variable weld seam profiles from the strong arc background based on intensity mutation and density feature detection for the thick plate welding process with T-joints. First, an improved Canny algorithm was used to magnify the weld seam profile and restrain interference. Next, an intensity mutation detection method was proposed to further strengthen the weld seam profile because there exists intensity mutation in the local region surrounding the weld seam profile. Finally, an algorithm based on the band-width and density feature scanning method was proposed to further eliminate the interference data after the strengthened image was binarized with the Otsu algorithm. The weld seam profile was identified as clusters with their spatial scale features after the nearest neighbor clustering was dealt with. The results show that this method can identify over 95% of the weld seam profiles from the arc interference background whose area is about 20% of the image. It provides valuable reference for promoting the automatic and intelligent welding process with different joints.

Nowadays, live broadcasting video has become increasingly popular and high‐quality live broadcasting video is highly needed. In practice, live broadcasting videos usually undergo several processing stages, which inevitably introduce multiple distortions, e. g. frame freezing and intensity mutation, causing the degraded quality of experience. However, little work has been done to the quality evaluation of live broadcasting videos, which may hinder the further development of more advanced live broadcasting video delivery systems. Motivated by this, this study presents a no‐reference quality evaluation model for live broadcasting videos (LBVQA) in temporal and spatial domains. In the temporal domain, statistic features are extracted to measure the frame freezing and intensity mutation, and the entropy‐based feature is extracted to describe the global jitter. In the spatial domain, blurring is measured based on phase coherence, and abnormal exposure ratio is calculated based on an adaptive threshold. Finally, all features are fed into a backpropagation neural network to train the quality prediction model. Experimental results on the Live Broadcasting Video Database demonstrate the advantages of the proposed metric over the state‐of‐the‐art image and video quality metrics. [ABSTRACT FROM AUTHOR]