Product Overview

The Vehicle Big-Data Platform (PVD) is a comprehensive big data combat platform built on the information-based demands for public security vehicle management. Supported by cloud computing, big data analysis, deep learning and other technologies, and centered on vehicle data, it realizes the network-based access, analysis and processing, cloud storage, smart study and judgment, visualized display of multi-source, massive vehicle resource information, and other functions. It effectively addresses such issues as the difficult access, few properties, difficult query and retrieval, and low-level deep application of massive vehicle data, and provides strong technical support for scientific and technical information, transportation, criminal investigation, command and dispatch, and other polices.

Technical Features

Based on the Power® Video Cloud solution, the optimization and parallel accelerated query engine technology, and vehicle recognition and smart algorithm in which NetPosa has proprietary intellectual property, the PVD of NetPosa realizes a full set of business processes including automatic snapshot, second-level query, automatic recognition, precise comparison and linked alarm through an integrated networking and integration of monitory gateway, electronic police and video resources. In this way, it truly realizes "retaining license plate number of vehicle, obtaining criminal evidence, depicting traces, and making in-depth study and judgment", and provides strong technical support for the police organs to investigate into and solve a case, crack down on and prevent crime.

Correct secondary extraction of vehicle information

Based on deep learning vehicle recognition algorithm, it achieves the secondary recognition of snapshot picture of front end monitory gateway, and can precisely recognize the license plate number, brand, sub-brand, year and model, color, license plate color, type, license plate type, annual inspection standards, sun shield, and other details of the vehicle, make up the deficiencies of front end monitory gateway in poor and inaccurate information recognition. Among others, vehicle type and license plate type support subdivision, and can recognize different years and models of the same vehicle type, generate the photos of annual inspection standards and other features, and offer more valuable vehicle information to deep application that is based on big data analysis.

Efficient retrieval capability

Based on the optimization and parallel accelerated query engine technology (OPAQ) with proprietary intellectual property, the platform realizes efficient retrieval capability, and the query speed can reach second-grade return of hundred-billion-grade passage record.

Deep data mining

Based on the big data analysis technology, it achieves the deep mining and application of massive data, and provides a complete study and judgment tool set. Over 30 kinds of strategies provide technical support for police users to study and judge the security condition and cramp down on vehicle-related illegal and criminal activities.

Open cloud architecture

The platform adopts the open cloud architecture, uploads vehicle license plate recognition and other smart algorithm, combines with the OPAQ in which NetPosa has proprietary intellectual property, supports parallel computing, use-on-demand, dynamic allocation and distributed deployment, and realizes more efficient storage, retrieval and computation.

Strong resources access capability

It realizes the connection with eight public security information resource pools. The access to public security resources and monitory gateway resources helps realize deep application, and achieves "searching person via vehicle" and "searching vehicle via person" through the three dimensions of person, vehicle and file.

Functions & Peculiarities

TIANWANG Vehicle Search

1. Monitory gateway monitoring

The equipment supports the retrieval and viewing of the real-time structuralized and historical structuralized results of video camera, the retrieval and viewing of the structuralized results of downloaded and uploaded video documents, and reverse image search. Based on the feature information of person, vehicle and non-motor vehicles, it helps users rapidly look for and locate suspected person or vehicle from massive videos.

3. Vehicle search by image

It supports the uploading of images, and can automatically recognize the entire vehicle information in the currently searched images, compares with the passage images. The retrieved results are ranked in the reverse order of similarity with the uploaded image, and passage data can be divided into groups based on license plate.

4. Vehicle search by features

It can extract and recognize partial feature information from the vehicle image with unknown license plate, so as to find vehicle same with the vehicle from the passage record.

Deep Mining

1. One vehicle one record

One vehicle one record: based on relevant technology of vehicle big data analysis, and on account of vehicle passage record, six-in-one platform information, personnel information, and information of case & event, among other information, it realizes record-based management of every vehicle, organizes and collects such resources as vehicle information, vehicle owner information, vehicle abnormality information, illegality information, and the retrieval and application status of vehicle, and assists the analysis, mining and application of the involved vehicle.

2. Wheel path

The system provides the function of analyzing the wheel path of designated vehicle, and allows users to designate vehicle license plate number and time range. The system will draw the dynamic wheel path on GIS based on the passage record in the searched time range.

Intelligent Study and Judgment

Based on smart video image analysis and big data technology, it provides over 30 strategies close to public security business.

1. Fake plate analysis

Fake plate query: supports the query of fake plate record within a period, and allows the query into whether a specific plate is a fake plate. As for a vehicle with a fake plate, it can further look into the basis of fake plate, judge whether the fake plate is an effective fake plate, analyze its real license plate of the vehicle, and produce the review report.

Fake-plate vehicle library: analyzes fake plate result and handling result within a period, demonstrates the same in the form of line graph, and displays with classified image-text concerning vehicle with effective fake plate and newly increased fake plate in the same day.

2. Analysis on driving without-a-seat-belt

It supports the analysis on vehicles where both the driver and co-pilot do not buckle up within a period, and supports the manual proofreading on whether the analysis result is valid or not when looking at the current vehicle details. It supports the export of retrieved results.

Search and Control

1. Blacklist management

It realizes the unified management and viewing of all blacklist control information imported into the platform, and supports screening and viewing by control status, control type, vehicle color and other information. It allows the classification and statistics of control information and manual uploading of the photo of the controlled vehicle, filling of the information of the controlled vehicle (precise license plate control and blurry license plate control), and the completion of entering of control information. In the meantime, it supports batch control, allows filling of excel table as per the template of control task, uploading of the same to the PVD, and completion of batch control. 

2. System management

It supports log management, user management, map labeling, system authorization, message center, task center, role management, operation and maintenance statistics, and routine query and analysis management of monitory gateway grouping function.