WO2024066129A1

WO2024066129A1 - Fault locating platform based on facial feature analysis

Info

Publication number: WO2024066129A1
Application number: PCT/CN2023/070625
Authority: WO
Inventors: 郑潇寒
Original assignee: 郑潇寒
Priority date: 2022-09-28
Filing date: 2023-01-05
Publication date: 2024-04-04

Abstract

The present invention relates to a fault locating platform based on facial feature analysis. The platform comprises: a data capturing device, which is arranged at a cloud server end and is used for capturing a picture that is currently acquired by a surveillance camera mechanism; an information analysis device, which is used for, on the basis of each multi-component value corresponding to each pixel point in the currently acquired picture and the total number of pixel units used by the surveillance camera mechanism, intelligently analyzing an aperture value, an ISO value and a shutter speed value used when the currently acquired picture is acquired; and an error identification device, which is used for determining whether a sensor for recording the aperture, the ISO or the shutter speed fails. By means of the present invention, a sensor fault parsing mechanism based on a feedforward neural network is designed at a cloud end, such that each camera parameter corresponding to picture content acquired in real time is intelligently parsed on the basis of the picture content acquired in real time and the total number of effective pixel units of a surveillance camera mechanism, thereby realizing intelligent determination on whether a local sensor of the surveillance camera mechanism fails.

Description

Fault location platform based on facial feature analysis

Technical Field

The present invention relates to the field of cloud computing services, and in particular to a fault location platform based on facial feature analysis.

Background technique

The relatively simple cloud computing technology has been widely used in today's Internet services, and the most common ones are network search engines and web mailboxes. The most familiar search engines are Google and Baidu. At any time, as long as you use a mobile terminal, you can search for any resources you want on the search engine and share data resources through the cloud. The same is true for web mailboxes. In the past, sending and writing an email was a relatively troublesome thing and a very slow process. Driven by cloud computing technology and network technology, emails have become a part of social life. As long as you are in a network environment, you can send and receive emails in real time. In fact, cloud computing technology has been integrated into today's social life.

Cloud computing is widely favored in various application fields because it transfers local computing to remote servers and realizes computing sharing at the same time. For example, it can be used in the field of visual monitoring. In the prior art, when each surveillance camera unit collects images, it will store the various camera parameters recorded by its various sensors, including aperture, ISO and shutter speed, together with the collected images to provide key data for subsequent image analysis and camera strategy selection. However, in actual use, the sensor data is incorrect due to design reasons or usage time reasons, and it cannot truly reflect the relevant camera environment, which seriously affects the judgment of the camera user and cannot realize the customized image collection operation for different environments and different collection requirements at the surveillance camera unit end.

Summary of the invention

In order to solve the above technical problems, the present invention provides a fault location platform based on facial feature analysis, which can design a targeted sensor fault analysis mechanism based on a feedforward neural network in the cloud, so as to intelligently analyze the various camera parameters corresponding to the picture content collected in real time by the monitoring camera mechanism based on the picture content collected in real time by the monitoring camera mechanism and the total number of effective pixel units of the monitoring camera mechanism, thereby providing valuable reference information for whether the local sensor of the monitoring camera mechanism has a fault.

According to one aspect of the present invention, a fault location platform based on facial feature analysis is provided, the platform comprising:

A data capture device is provided at a cloud server end connected to the network of the surveillance camera mechanism, and is used to capture the picture currently collected by the surveillance camera mechanism as an instant captured picture;

A content conversion device, arranged at the cloud server end and connected to the data capture device, is used to obtain a multi-component value corresponding to each pixel point in the received instant captured picture;

A parameter extraction device, which is arranged on the cloud server side and connected to the monitoring camera mechanism, is used to obtain the total number of pixel units of the image sensor of the monitoring camera mechanism that are in a working state when collecting the instant captured picture, and output it as the number of collection units;

An information analysis device, arranged at the cloud server end and connected to the content conversion device and the parameter extraction device respectively, for intelligently analyzing the aperture value, ISO value and shutter speed value of the monitoring camera mechanism when the real-time captured picture is collected based on the multi-component values corresponding to the pixels in the real-time captured picture and the number of the acquisition units, and outputting them as the identification aperture value, identification ISO value and identification shutter speed value respectively;

A value judgment device is arranged inside the monitoring camera mechanism, and is used to obtain the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded by the monitoring camera mechanism when the monitoring camera mechanism acquires the instant captured picture;

An error identification device is arranged on the cloud server end and is respectively connected to the value judgment device and the information analysis device, and is used to judge that a sensor for recording the aperture value, ISO value or shutter speed value of the monitoring camera mechanism is faulty and send a sensor fault signal when the error between the aperture value and the collected aperture value exceeds a limit, the error between the ISO value and the collected ISO value exceeds a limit, or the error between the shutter speed value and the collected shutter speed value exceeds a limit.

The present invention has at least the following significant technical effects:

First, a targeted sensor fault analysis mechanism is designed in the cloud to perform intelligent analysis operations on the sensor faults recorded by the aperture, ISO, and shutter speed of each network-connected surveillance camera.

Secondly, in the specific intelligent analysis, the multi-component values corresponding to each pixel point in the real-time captured image captured by the surveillance camera mechanism and the number of effective pixel units actually used by the surveillance camera mechanism to capture the real-time captured image are used as input data of the intelligent analysis model to obtain the aperture value, ISO value and shutter speed value of the surveillance camera mechanism, thereby providing reliable parameter data for the identification of sensor failures of the surveillance camera mechanism.

BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

FIG1 is a schematic diagram of the structural topology of a feedforward neural network used in a fault location platform based on facial feature analysis according to various embodiments of the present invention.

FIG2 is a schematic diagram of the structure of a fault location platform based on facial feature analysis according to Embodiment A of the present invention.

FIG3 is a schematic structural diagram of a fault location platform based on facial feature analysis according to Embodiment B of the present invention.

Detailed ways

The implementation scheme of the fault location platform based on facial feature analysis of the present invention will be described in detail below with reference to the accompanying drawings.

Example 1

As shown in FIG1 , the feedforward neural network includes a plurality of inputs X1 , X2 , X3 and X4 and a plurality of outputs Y1 , Y2 and Y3 , and the feedforward neural network includes an input layer, a hidden layer and an output layer.

In Figure 1, the multiple inputs of the feedforward neural network are the multi-component values and the number of acquisition units corresponding to each pixel point in the real-time captured picture, and the multiple outputs of the feedforward neural network are the aperture value, ISO value and shutter speed value of the monitoring camera mechanism when collecting the real-time captured picture.

Example 2

FIG2 is a schematic diagram of the structure of a fault location platform based on facial feature analysis according to Embodiment A of the present invention, wherein N is a natural number greater than 1, and the platform includes:

An information analysis device, arranged at the cloud server end and connected to the content conversion device and the parameter extraction device respectively, for intelligently analyzing the aperture value, ISO value and shutter speed value of the monitoring camera mechanism when collecting the real-time captured picture based on the multi-component values corresponding to the pixels in the real-time captured picture and the number of the acquisition units, and outputting them as the identification aperture value, identification ISO value and identification shutter speed value respectively;

an error identification device, arranged at the cloud server end and connected to the value judgment device and the information analysis device respectively, for judging that a sensor for recording the aperture value, ISO value or shutter speed value of the monitoring camera mechanism is faulty and sending a sensor fault signal when an error between the aperture value and the collected aperture value exceeds a limit, an error between the ISO value and the collected ISO value exceeds a limit, or an error between the shutter speed value and the collected shutter speed value exceeds a limit;

Wherein, capturing the picture currently collected by the monitoring camera mechanism as the real-time captured picture includes: the monitoring camera mechanism includes a resolution adjustment unit, an imaging lens, an optical lens component and a photoelectric sensor component;

And judging that a sensor for collecting aperture value, collecting ISO value or collecting shutter speed value of the monitoring camera mechanism fails and sending a sensor failure signal includes: sending the sensor failure signal by using a light emitting action with a set frequency.

Example 3

Different from the implementation scheme A in FIG2 , the fault location platform based on facial feature analysis in the implementation scheme B may further include:

A touch display device, arranged on the front screen of the monitoring camera mechanism, connected to the error identification device, and used for receiving and displaying in real time a warning text message related to the sensor fault signal or the sensor error-free signal;

Wherein, obtaining the multi-component value corresponding to each pixel point in the received instant captured image includes: the multi-component value corresponding to each pixel point includes the hue component value, brightness component value and saturation component value corresponding to the pixel point in the HSV space;

Among them, obtaining the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded by the surveillance camera mechanism when the surveillance camera mechanism captures the real-time captured image includes: obtaining the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded respectively by each sensor of the surveillance camera mechanism when the surveillance camera mechanism captures the real-time captured image.

Next, the specific structure of the fault location platform based on facial feature analysis of the present invention will be further described.

In the fault location platform based on facial feature analysis according to various embodiments of the present invention:

The error identification device is also used to determine that each sensor of the monitoring camera mechanism for collecting aperture value, collecting ISO value and collecting shutter speed value records is not faulty and send a sensor error-free signal when the error between the identified aperture value and the collected aperture value is within the limit, the error between the identified ISO value and the collected ISO value is within the limit, and the error between the identified shutter speed value and the collected shutter speed value is within the limit;

Wherein, when the error between the identified aperture value and the collected aperture value exceeds the limit, the error between the identified ISO value and the collected ISO value exceeds the limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, judging that the sensor for collecting the aperture value, collecting the ISO value, or collecting the shutter speed value of the monitoring camera mechanism is faulty and sending a sensor fault signal includes: when the error between the identified aperture value and the collected aperture value exceeds the limit, judging that the sensor for collecting the aperture value of the monitoring camera mechanism is faulty and sending a sensor fault signal;

Among them, when the error between the identified aperture value and the collected aperture value exceeds the limit, the error between the identified ISO value and the collected ISO value exceeds the limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, it is judged that the sensor for recording the collected aperture value, the collected ISO value, or the collected shutter speed value of the monitoring camera mechanism is faulty and a sensor fault signal is issued, including: when the error between the identified ISO value and the collected ISO value exceeds the limit, it is judged that the sensor for recording the collected ISO value of the monitoring camera mechanism is faulty and a sensor fault signal is issued.

Based on the multi-component values corresponding to the pixels in the real-time captured image and the number of the acquisition units, the aperture value, ISO value and shutter speed value of the surveillance camera mechanism when the real-time captured image is captured are intelligently analyzed and collected and output as the identification aperture value, identification ISO value and identification shutter speed value respectively, including: using the multi-component values corresponding to the pixels in the real-time captured image and the number of the acquisition units as two input data of the feedforward neural network to run the feedforward neural network and obtain the identification aperture value, identification ISO value and identification shutter speed value;

Wherein, using the multi-component values corresponding to the pixels in the instant captured image and the number of acquisition units as two input data of the feedforward neural network to run the feedforward neural network and obtain the identification aperture value, the identification ISO value and the identification shutter speed value includes: performing normalization processing on the multi-component values corresponding to the pixels in the instant captured image and the number of acquisition units and inputting them into the feedforward neural network;

The method further comprises: using the multi-component values corresponding to the pixels in the instant captured image and the number of the acquisition units as two input data of the feedforward neural network to run the feedforward neural network and obtain the identification aperture value, the identification ISO value and the identification shutter speed value; the obtained identification aperture value, the identification ISO value and the identification shutter speed value are all in a hexadecimal numerical representation mode;

Among them, the multi-component values corresponding to each pixel point in the real-time captured picture and the number of acquisition units are respectively normalized and input into the feedforward neural network, including: the multi-component values corresponding to each pixel point in the real-time captured picture and the number of acquisition units are respectively hexadecimal encoded and input into the feedforward neural network.

In addition, in the fault location platform based on facial feature analysis, when the error between the identified aperture value and the collected aperture value exceeds the limit, the error between the identified ISO value and the collected ISO value exceeds the limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, it is judged that the sensor for recording the collected aperture value, the collected ISO value, or the collected shutter speed value of the monitoring camera mechanism has a fault and sends a sensor fault signal, including: when the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, it is judged that the sensor for recording the collected shutter speed value of the monitoring camera mechanism has a fault and sends a sensor fault signal.

The fault location platform based on facial feature analysis of the present invention is used to address the technical problem in the prior art of lacking a real-time judgment mechanism for whether various types of sensors of each surveillance camera mechanism have faults. By designing a sensor fault analysis mechanism based on a feedforward neural network in the cloud, the various camera parameters corresponding to the real-time collected image content are intelligently analyzed based on the real-time collected image content and the total number of effective pixel units of the surveillance camera mechanism, thereby realizing intelligent judgment on whether the local sensor of the surveillance camera mechanism has faults.

In addition, the embodiment of the present invention is not limited to the above-mentioned embodiment, and various changes can be made without departing from the scope of the present invention.

Claims

A fault location platform based on facial feature analysis, characterized in that the platform comprises:

A data capture device is provided at a cloud server end connected to the network of the surveillance camera mechanism, and is used to capture the picture currently collected by the surveillance camera mechanism as an instant captured picture;

A content conversion device, arranged at the cloud server end and connected to the data capture device, is used to obtain a multi-component value corresponding to each pixel point in the received instant captured picture;

A parameter extraction device, which is arranged on the cloud server side and connected to the monitoring camera mechanism, is used to obtain the total number of pixel units of the image sensor of the monitoring camera mechanism that are in a working state when collecting the instant captured picture, and output it as the number of collection units;

An information analysis device, arranged at the cloud server end and connected to the content conversion device and the parameter extraction device respectively, for intelligently analyzing the aperture value, ISO value and shutter speed value of the monitoring camera mechanism when collecting the real-time captured picture based on the multi-component values corresponding to the pixels in the real-time captured picture and the number of the acquisition units, and outputting them as the identification aperture value, identification ISO value and identification shutter speed value respectively;

A value judgment device is arranged inside the monitoring camera mechanism, and is used to obtain the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded by the monitoring camera mechanism when the monitoring camera mechanism acquires the instant captured picture;

An error identification device is arranged on the cloud server side and is respectively connected to the value judgment device and the information analysis device, and is used to judge that a sensor for recording the aperture value, ISO value or shutter speed value of the monitoring camera mechanism is faulty and send a sensor fault signal when the error between the identified aperture value and the collected aperture value exceeds a limit, the error between the identified ISO value and the collected ISO value exceeds a limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds a limit.
The fault location platform based on facial feature analysis according to claim 1, characterized in that the platform also includes:

A touch display device, arranged on the front screen of the monitoring camera mechanism, connected to the error identification device, and used for receiving and displaying in real time a warning text message related to the sensor fault signal or the sensor error-free signal;

Wherein, obtaining the multi-component value corresponding to each pixel point in the received instant captured image includes: the multi-component value corresponding to each pixel point includes the hue component value, brightness component value and saturation component value corresponding to the pixel point in the HSV space;

Among them, obtaining the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded by the surveillance camera mechanism when the surveillance camera mechanism captures the real-time captured image includes: obtaining the acquisition aperture value, acquisition ISO value and acquisition shutter speed value recorded respectively by each sensor of the surveillance camera mechanism when the surveillance camera mechanism captures the real-time captured image.
The fault location platform based on facial feature analysis according to any one of claims 1-2, characterized in that:

The error identification device is also used to determine that the sensors of the monitoring camera mechanism that perform the collection of aperture values, ISO values, and shutter speed value records are not faulty and send a sensor error signal when the error between the identified aperture value and the collected aperture value is within the limit, the error between the identified ISO value and the collected ISO value is within the limit, and the error between the identified shutter speed value and the collected shutter speed value is within the limit.
The fault location platform based on facial feature analysis as claimed in claim 3, characterized in that:

When the error between the identified aperture value and the collected aperture value exceeds the limit, the error between the identified ISO value and the collected ISO value exceeds the limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, it is determined that the sensor for recording the collected aperture value, the collected ISO value, or the collected shutter speed value of the monitoring camera mechanism is faulty and a sensor fault signal is issued, including: when the error between the identified aperture value and the collected aperture value exceeds the limit, it is determined that the sensor for recording the collected aperture value of the monitoring camera mechanism is faulty and a sensor fault signal is issued.
The fault location platform based on facial feature analysis as claimed in claim 4, characterized in that:

When the error between the identified aperture value and the collected aperture value exceeds the limit, the error between the identified ISO value and the collected ISO value exceeds the limit, or the error between the identified shutter speed value and the collected shutter speed value exceeds the limit, it is determined that the sensor for recording the collected aperture value, the collected ISO value, or the collected shutter speed value of the monitoring camera mechanism is faulty and a sensor fault signal is issued, including: when the error between the identified ISO value and the collected ISO value exceeds the limit, it is determined that the sensor for recording the collected ISO value of the monitoring camera mechanism is faulty and a sensor fault signal is issued.
The fault location platform based on facial feature analysis according to any one of claims 1-2, characterized in that:

Based on the multi-component values corresponding to each pixel point in the real-time captured picture and the number of the acquisition units, the aperture value, ISO value and shutter speed value of the monitoring camera mechanism when the real-time captured picture is captured are intelligently analyzed and output as the identification aperture value, identification ISO value and identification shutter speed value respectively, including: using the multi-component values corresponding to each pixel point in the real-time captured picture and the number of the acquisition units as two input data of the feedforward neural network to run the feedforward neural network and obtain the identification aperture value, identification ISO value and identification shutter speed value.
The fault location platform based on facial feature analysis according to claim 6, characterized in that:

Using the multi-component values corresponding to each pixel point in the real-time captured image and the number of acquisition units as two input data of a feedforward neural network to run the feedforward neural network and obtain the identified aperture value, the identified ISO value and the identified shutter speed value includes: performing normalization processing on the multi-component values corresponding to each pixel point in the real-time captured image and the number of acquisition units and then inputting them into the feedforward neural network.
The fault location platform based on facial feature analysis according to claim 7, characterized in that:

Using the multi-component values corresponding to each pixel point in the instant captured image and the number of acquisition units as two input data of the feedforward neural network to run the feedforward neural network and obtain the identification aperture value, identification ISO value and identification shutter speed value also includes: the obtained identification aperture value, identification ISO value and identification shutter speed value are all in hexadecimal numerical representation mode.
The fault location platform based on facial feature analysis according to claim 8, characterized in that:

The method of normalizing the multi-component values corresponding to each pixel point in the real-time captured image and the number of acquisition units and inputting them into the feedforward neural network includes: performing hexadecimal encoding on the multi-component values corresponding to each pixel point in the real-time captured image and the number of acquisition units and inputting them into the feedforward neural network.