A method and device for target detection

Abstract

The target detection method, device and monitoring system provided by the present invention, based on the improved target detection method of background difference, divide the background image and the image to be tested into several areas, and determine whether the captured area has changed according to the change of the image information value of each area, There is no need to extract the exact outline of the detected target, reducing the amount of calculation; calculate the target confirmation value according to the changed area parameters in the image to be tested; and judge whether the abnormal situation in the photographed area is a monitoring target according to the target confirmation value. The interference effect of objective factors such as scene light on determining the monitoring target is eliminated, and the accuracy of target detection is improved.

Description

technical field [0001] The invention relates to monitoring technology, in particular to a target detection method and device. Background technique [0002] In the aspect of moving target detection, it mainly includes inter-frame difference method, method based on background difference and method based on optical flow field. Among them, the inter-frame difference method has a small amount of calculation, fast detection speed, and is easy to implement, but it is also easily affected by noise and is more sensitive to the influence of objective factors such as weather and light. If the illumination changes or the scene is rainy and snowy, it is easy to detect Wrong; the background-based method needs to perform threshold segmentation and binarize the image. At the same time, in order to eliminate the influence of the rough edge contour and noise of the detection target, it needs to perform morphological filtering on the image, which requires a large amount of calculation. It is ...

AI Technical Summary

Problems solved by technology

Among them, the inter-frame difference method has a small amount of calculation, fast detection speed, and is easy to implement, but it is also easily affected by noise and is more sensitive to the influence of objective factors such as weather and light. If the illumination changes or the scene is rainy and snowy, it is easy to detect Wrong; the background-based method needs to perform threshold segmentation and binarize the image. At the same time, in order to eliminate the influence of the rough edge contour and noise of the detection target, it needs to perform morphological filtering on the image, which requires a large amount of calculation. It is impossible to eliminate the impact of sudden changes in the scene and illumination; the advantage of the method based on the optical flow field is that it can detect independent moving objects, does not need to know any information about the scene in advance, and can be used in the case of camera movement, but this The calculation of the method is complex and time-consuming. To achieve real-time detection, the hardware requirements are very high.

[0003] In addition, in the existing monitoring technology, the existence of objective factors increases the difficulty of monitoring a specific area, and the monitoring system is prone to false alarms

For example, the monitoring system used for home monitoring may become the target of monitoring because pets move in the monitoring area. When the monitoring system detects pet movement, it will also enter the alarm state and cause false alarms.

Images