What is the intelligent video analysis camera intelligent video analysis method

Intelligent video analysis camera is a branch of computer image vision technology applied in security field, and it is an intelligent monitoring technology based on target behavior. Different from the traditional VMD (Video Motion Detection) technology, the intelligent video analysis camera first separates the background and the target in the scene, recognizes the real target, and removes background interference (such as leaf shake, surface wave, light change). Then analyze and track the target behavior that occurs within the camera scene. The biggest difference between intelligent video analysis cameras and other smart cameras is the intelligent video analysis. Today, Xiaobian will introduce you to the analysis method of intelligent video analysis cameras.

There are two main types of video analysis methods for intelligent video analysis cameras:

1, background subtraction method

A method of detecting a motion region using the difference between the current image and the background image (SAD). More complete moving target feature data can be provided. High precision and sensitivity, with good performance. Modeling of the background is the technical key to background subtraction. Generally, the system adaptive learning time is set in the system setting period to model, and the learning time of 3-5 minutes is selected according to the actual "live level" of the background. After the general system modeling is completed, the background will change with time. The system has the ability of “background maintenance”, that is, some images that are later integrated into the background, such as clouds, can be automatically added as background.

2, time difference method

The time difference method is an advanced VMD, also known as the adjacent frame difference method, which uses the video image feature to extract the required dynamic target information from the continuously obtained video stream. The time difference is essentially the use of subtraction of adjacent frame images to extract information about the movement of the foreground object. This method can not completely extract all the relevant feature pixel points, and may generate voids inside the motion entity, and intelligently detect the edge of the target.

The above is about the introduction of intelligent video analysis cameras. Intelligent video analysis cameras are widely used in public safety related systems, building intelligence, intelligent transportation and other related fields.