Seaview IV007 Video Structured System

Seaview IV007 Video Structured System Target detection, object recognition and in-depth learning are used to extract target objects and trajectories in video images. Target objects are classified into different categories, such as human, bicycle, human cycling, behavior and events. High-level structured attributes of targets are further extracted, including target color features and classification features. Velocity features, etc., and the target trajectory information and high-level features form an efficient index data for structured retrieval.

Seaview IV007 Video Structured System

I. Core Functions

Seaview IV007 Video Structured SystemFor the people, vehicles and non-motor vehicle categories which are particularly concerned in video images, it provides deeper structural analysis. For the characters in the video image, face detection and face recognition can be used to analyze the specific identity of the characters, and various structured attributes of pedestrians can be provided, including attributes of clothing and decorations: jacket, trousers, skirts and dresses, shoes, hats, sunglasses, shawls, scarves, belts and waistbands; Carrier characteristics: single shoulder bag, double shoulder backpack, handbag, pole box, umbrella; human characteristics: hair, face.

For vehicles in video images, multi-lane vehicle detection, vehicle front and rear detection and recognition functions can be carried out, and more than 10 structured attribute information can be extracted to identify vehicles, includingVehicle number plate, body color, vehicle brand, vehicle type, sub-brand, vehicle year and various vehicle characteristics information, such as: annual inspection, sunshade, pendant, pendant, paper towel box, seat belt, etc.

Structured cycling is the structured processing and recognition of cyclists in video resources. It includes the structured processing of cyclists'clothing type, accessories, body shape, vehicle color, orientation, number of bikes, umbrellas and so on, which is related to the external characteristics of cycling.

Target Trajectory Extraction

After structured recognition of a pedestrian, motor vehicle and bicycle target object, the trajectory of the target object can be quickly extracted, which saves search time and determines the target object's trajectory with one key.

II. Structural Capability

Seaview IV007 Video Structured SystemWith in-depth learning and artificial intelligence technology, various attributes of human, vehicle and non-motor vehicle are automatically identified and extracted. It has the most abundant ability to extract target attributes in the industry. It provides strong technical support for the use of large video data. The main structured data information is as follows.

III. Application scenarios

1. Real-time video stream structured analysis, system based on structured analysis results for alarm, retrieval and mining.

2. Structural analysis of offline video file upload. The system carries out large data analysis and retrieval based on the analysis results.

3. Docking video networking platform, access large-scale real-time and offline video, unify video structured analysis, and build a deep application platform for large-scale video data.