Analyzing scene image features with adaptive nonlinear diffusion equation filters for public safety project scheduling

Abstract

Abstract The current image feature detection methods fail to perform image edge filtering in all directions, leading to target deviation of the feature detection results, and narrower brightness level range of gray images. Hence, a novel approach based on adaptive nonlinear diffusion equation filtering is proposed for detecting security events in public safety project scheduling. An edge detection operator is developed that can work in any direction, and adaptive nonlinear diffusion image edge morphological composite filtering is implemented using lag threshold and NMS. The shape response of the image edge is calculated to select the adaptive threshold of feature extraction. A gaussian function pyramid is constructed to obtain the feature distribution sequence considering multi-scale factors. The wavelet analysis is used to analyze the gray scale and detail features of the scene images of public security events, and the gray projection method is utilized to detect the target features of dynamic images, completing the scene image feature detection of public security events. The experimental results demonstrate that the proposed feature detection method has high accuracy, shorter time, and ideal application performance.