Search

As most anomaly detection methods generally ignore the impact of video quality fluctuations in actual scenarios on their performance. Firstly, public datasets from the field of video anomaly detection were selected, and a distortion dataset for anomaly detection by artificially introducing common distortions found in surveillance videos through simulation methods was constructed. This paper systematically analyzes the impact of video quality fluctuations on the performance of four classic video anomaly detection algorithms. And the experimental results show that the degradation of surveillance data quality have a negative effect on the effectiveness of the video anomaly detection algorithm. Moreover, the impact of different distortion types and different degrees of distortion on the algorithm is different, as well as, Gaussian noise distortion and over brightness distortion have a greater impact on the algorithm. [ABSTRACT FROM AUTHOR]

As a food consumed worldwide, ginger is often sulfur-fumigated. Sulfur-fumigated ginger is harmful to health. However, traditional methods to detect sulfur-fumigated ginger are expensive and unpractical for the general public. In this paper, we present an efficient and convenient identification method based on image processing. First, rapid detection kits were employed to mark three levels of sulfur-fumigated gingers, and the RGB images of the gingers of each sulfur-fumigated level are collected. Second, the brightness and texture features were extracted from the images. Three machine learning methods, Support Vector Machine, Back Propagation Neural Network and Random Forest, were applied to establish prediction models. Third, the accuracy of each model was calculated and different weights were assigned for different models. Finally, models with different weights determined whether the ginger was sulfur-fumigated or non-sulfur-fumigated, and then the results were summarized to establish the final identification model. The experimental results show that the proposed method is robust. When the training set occupies 90%, the prediction accuracy is up to 100%. When the training set only occupies 10%, the accuracy remains high at 80%. Meanwhile, the proposed method is more competitive than other methods in terms of accuracy. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Henan Agricultural Sciences is the property of Editorial Board of Journal of Henan Agricultural Sciences and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)