Method, device and application for detection of garbage dumping in violation of regulations

Abstract

This application proposes a method, device and application for detecting illegal littering behavior, including the steps of: acquiring continuous multi-frame monitoring video images; acquiring the monitoring video images in which at least one garbage detection frame and at least one human body detection frame exist at the same time to obtain the first A set of images to be detected; obtaining any first image to be detected in the first set of images to be detected, obtaining a garbage feature map of each of the garbage detection frames of the first image to be detected and each of the human bodies Detecting the human body feature map of the frame; obtaining the interaction features between each of the garbage feature maps and each of the human body feature maps in the first image to be detected; classifying the interaction features to obtain a classification result, based on the classification As a result, the illegal littering behavior was determined. This scheme can improve the detection accuracy of the illegal littering behavior. At the same time, it can pay more attention to garbage and littering personnel in a complex detection environment, and reduce the detection pressure of the detection algorithm.

Description

technical field [0001] This application relates to the field of behavior monitoring, in particular to a method, device and application for detecting behavior of discarding garbage in violation of regulations. Background technique [0002] The comprehensive improvement of urban environment is a management project that relies on the division of labor and cooperation of various urban management departments, uses various technical means, organizes and supervises the joint efforts of various units and citizens to prevent and control environmental pollution. In the comprehensive improvement project of urban environment, the behavior of discarding garbage in violation of regulations is an uncivilized behavior that needs to be rectified urgently. The result of illegal dumping of garbage is that all kinds of domestic garbage are piled up in public places in an orderly manner, and such behavior will cause various social problems: for example, if the garbage exposed to the air is not c...

AI Technical Summary

Problems solved by technology

The result of illegal dumping of garbage is that all kinds of domestic garbage are piled up in public places in an orderly manner, and such behavior will cause various social problems: for example, if the garbage exposed to the air is not cleaned up in time, It will lead to the breeding of mosquitoes and flies, which will infect diseases and endanger human health; the garbage discarded into the water source will cause the death of marine animals, which will affect the sanitation of drinking water; in addition, random piles of garbage will also bring great safety Potential hazards, these garbage can easily cause fire when exposed to open flames

[0003] At present, there are two types of measures to supervise illegal dumping of garbage: manual supervision and video supervision. The manual supervision method refers to the regular inspection of specific places by full-time urban management personnel. This method is time-consuming and labor-intensive, with small coverage and poor timeliness. problems; video supervision refers to the real-time supervision of supervision places with the help of computer vision technology, so as to automatically supervise illegal littering behaviors. The method can't realize the detection of illegal littering behavior very well and accurately.

Specifically, most of the current illegal littering behavior detection models use the behavior characteristics of littering to detect littering behavior directly, but since there are often multiple people and objects in the background of the monitoring scene, multiple people and There are various behavior patterns between multiple objects. The current detection model for illegal littering behavior needs to traverse all the background factors and make behavioral judgments one by one to get the detection results, resulting in a huge computational pressure for the model; and, the current violation The littering behavior detection model judges the behavior characteristics of littering and pays more attention to the relative position between garbage and people. This method is difficult to distinguish the behavioral characteristics between illegal littering behavior and normal behavior, which leads to the illegal littering behavior detection model. The test results are difficult to meet the actual urban management needs

Images