An all-weather automatic inspection method and system for illegal activities

Abstract

The present invention provides an all-weather automatic inspection method and system for illegal activities, which includes the following steps: S1, determine the warning area; S2, acquire the image of the area to be inspected, and identify the target to be inspected; S3, if it is determined that the inspection area If the target is within the warning area, it is determined that the target to be detected for the first time is the first intrusion target; at the same time, the attitude information of the imaging device is recorded and stored in the illegal intrusion database; S4, judging whether the first intrusion target exists other illegal acts; and S5, sequentially identify the target to be checked frame by frame on the remaining frames of the image of the area to be checked, and determine whether the intrusion target is a target that appears in the warning area for the first time. The present invention can acquire clear image information of a small target in a long-distance scene, thereby realizing accurate detection of the small target, improving detection accuracy, reducing detection tasks, and improving detection efficiency.

Description

technical field [0001] The invention relates to the technical field of unmanned aerial vehicles, in particular to an all-weather automatic inspection method and system for illegal activities. Background technique [0002] Real-time detection and evidence collection of illegal activities in specific scenarios has always been a pain point for relevant departments. Current routine testing methods include: [0003] 1. Manual inspection requires a lot of labor costs, and with the increase of working hours, there is a certain instability in the standard of manual judgment; [0004] 2. Use high-point fixed cameras to monitor the designated warning area, and then use the target detection algorithm to detect pedestrians, ships and other targets, and send them to the command platform to judge whether the target is an illegal target according to whether it is in the warning area. The camera cannot detect illegal activities at night, and cannot achieve all-weather monitoring. At the s...

AI Technical Summary

Problems solved by technology

[0003] 1. Manual inspection requires a lot of labor costs, and with the increase of working hours, there is a certain instability in the standard of manual judgment;

[0004] 2. Use high-point fixed cameras to monitor the designated warning area, and then use the target detection algorithm to detect pedestrians, ships and other targets, and send them to the command platform to judge whether the target is an illegal target according to whether it is in the warning area. The camera cannot detect illegal activities at night, and cannot achieve all-weather monitoring. At the same time, in long-distance scenarios, the target initially recognized by the camera may be too small, resulting in unclear detected targets and inability to obtain evidence. Confirmation, resulting in low detection efficiency, unable to quickly capture illegal behaviors and conduct judgment and evidence collection

Images