Detection method and system based on Vibe algorithm and artificial neural network

• Main Authors: QIU SHUHONG, GAN HAIHUA, JIANG MINGYUAN, TONG SUN
• Format: Patent
• Language: Chinese; English
• Published: 26.08.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention relates to the field of computer vision, in particular to a detection method and system based on a Vibe algorithm and an artificial neural network. According to the method, the moving target of the monitoring image is extracted through the Vibe background modeling algorithm, the situation that the small-size moving target is omitted can be avoided, and the accuracy of detecting the moving target can be greatly improved. In addition, the low-resolution image is reconstructed through the generative adversarial network, and the resolution of the moving target image is improved. On the basis that the image resolution is improved, the accuracy of detecting the moving target is remarkably improved. And finally, the loss of the ESRGAN is replaced by Wassersteinloss, so that the training process of the generative adversarial network is smoother and more stable. And the classification effect of the neural network classifier is remarkably improved by improving Arcfieloss. The performance optimization of t