Visual surveillance image content identification equipment

Abstract

The present invention relates to a visual monitoring image content identification device, comprising: a rear-end monitoring component fixed at the rear end of an electric vehicle for performing visual monitoring of the rear-end scene of the running electric vehicle, and for acquiring and outputting a corresponding on-site image signal; a multi-level processing mechanism for intelligently identifying whether the on-site image signal is a BAYER image directly output by a photoelectric sensor of the rear-end monitoring component based on multiple visual data of the on-site image signal and the width of the driving road of the rear-end scene of the electric vehicle and the nearest car distance. Through the present invention, a multi-level processing mechanism with a customized hardware architecture and a targeted artificial intelligence model can be introduced to perform intelligent identification of whether the on-site image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component, thereby avoiding misjudgment in the identification and early warning of rear-end collisions and other safety accidents.

Description

Technical Field

[0001] The present invention relates to the field of visual monitoring, and in particular to a visual monitoring image content identification device. Background Art

[0002] Compared with fuel vehicles, electric vehicles mainly differ in four major components: drive motor, speed controller, power battery, and on-board charger. Compared with gas stations, they have public ultra-fast charging stations. The quality difference of electric vehicles depends on these four major components, and their value also depends on the quality of these four major components. The purpose of electric vehicles is also directly related to the selection and configuration of the four major components. The speed and starting speed of electric vehicles depend on the power and performance of the drive motor, and the length of their mileage depends on the capacity of the on-board power battery. The weight of the on-board power battery depends on the type of power battery used, such as lead-acid, zinc-carbon, lithium battery, etc. They have different volumes, specific gravity, specific power, specific energy, and cycle life. This depends on the manufacturer's positioning and use of the vehicle grade, as well as market definition and market segmentation.

[0003] During the driving process of electric vehicles, due to the flammability of power batteries, the monitoring of scene safety is better than that of fuel vehicles. For example, a back-end monitoring component is needed to perform visual monitoring of the back-end scene of the running electric vehicle, to obtain and output the corresponding on-site image signal, thereby providing basic data for the identification and early warning of possible rear-end collisions and other safety accidents. It can be seen that whether the on-site image signal can completely restore the back-end scene of the electric vehicle is crucial. Some image processing or special effects processing that the on-site image signal has undergone may affect the identification and early warning of the above-mentioned phenomena or accidents. Summary of the invention

[0004] In order to solve the technical problems in the related fields, the present invention provides a visual monitoring image content identification device, which can introduce a multi-level processing mechanism with a customized hardware architecture including a primary processing component, a secondary processing component and a final processing component as a hardware resource for intelligently identifying whether the on-site image signal is a BAYER image directly output by the photoelectric sensor of the back-end monitoring component. A targeted artificial intelligence model is used to perform intelligent identification of whether the on-site image signal is a BAYER image directly output by the photoelectric sensor of the back-end monitoring component, thereby ensuring the stability and reliability of the intelligent identification results, and providing more effective basic data for the identification and early warning of possible rear-end collisions and other safety accidents.

[0005] According to the present invention, a visual surveillance image content identification device is provided, the device comprising:

[0006] A rear-end monitoring component, fixed at the rear end of the electric vehicle, is used to perform visual monitoring of the rear-end scene of the running electric vehicle, and is used to acquire and output corresponding on-site image signals;

[0007] A target reshaping component is used to reshape the deep neural network to obtain an intelligent analysis body, wherein the reshaping of the deep neural network to obtain the intelligent analysis body includes: performing training operations on the deep neural network to obtain the deep neural network after completing each training operation and serving as the intelligent analysis body;

[0008] A multi-stage processing mechanism, connected to the rear-end monitoring component and the target reshaping component respectively and comprising a primary processing component, a secondary processing component and a final processing component, wherein the primary processing component is used to extract the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the secondary processing component is used to obtain the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle, and the final processing component is connected to the secondary processing component and the primary processing component respectively, and is used to intelligently identify whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle;

[0009] An information transmission mechanism, connected to the intelligent analysis device, for transmitting the received intelligent identification results to a remote big data server through a network connection channel;

[0010] Among them, the BAYER image that intelligently identifies whether the scene image signal is the direct output of the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types, and each brightness value corresponding to each pixel of the scene image signal, as well as the width of the driving road of the rear-end scene of the electric vehicle and the nearest car distance includes: using an intelligent analysis body to intelligently identify whether the scene image signal is the direct output of the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types, and each brightness value corresponding to each pixel of the scene image signal, as well as the width of the driving road of the rear-end scene of the electric vehicle and the nearest car distance.

[0011] Therefore, the present invention has at least the following three important invention points:

[0012] Invention point 1: A multi-level processing mechanism with a customized hardware architecture is introduced as a hardware resource for intelligently identifying whether a scene image signal is a BAYER image directly output by a photoelectric sensor of a back-end monitoring component. Specifically, the multi-level processing mechanism includes a primary processing component, a secondary processing component, and a final processing component;

[0013] Invention point 2: The structure of the artificial intelligence model for intelligent identification of whether the on-site image signal is the BAYER image directly output by the photoelectric sensor of the back-end monitoring component is a targeted structure, thereby ensuring the stability and reliability of the intelligent identification result, wherein each training operation is performed on the deep neural network to obtain the deep neural network after each training operation and serve as an intelligent analysis body, and the number of times the deep neural network is trained is proportional to the number of pixels constituting the on-site image signal output by the back-end monitoring component;

[0014] Invention point three: The contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types of the on-site image signal and the brightness values ​​corresponding to the various pixels constituting the on-site image signal and various scene information of the rear-end scene of the electric vehicle are specifically screened and used as the basic data for intelligent identification, thereby further ensuring the stability and reliability of the intelligent identification results, wherein the various scene information of the rear-end scene of the electric vehicle are the width of the driving road of the rear-end scene of the electric vehicle and the distance to the nearest car. DETAILED DESCRIPTION

[0015] The following is a detailed description of an embodiment of the visual surveillance image content identification device of the present invention.

[0016] The visual surveillance image content identification device shown in embodiment A of the present invention includes:

[0017] A rear-end monitoring component, fixed at the rear end of the electric vehicle, is used to perform visual monitoring of the rear-end scene of the running electric vehicle, and is used to acquire and output corresponding on-site image signals;

[0018] A target reshaping component is used to reshape the deep neural network to obtain an intelligent analysis body, wherein the reshaping of the deep neural network to obtain the intelligent analysis body includes: performing training operations on the deep neural network to obtain the deep neural network after completing each training operation and serving as the intelligent analysis body;

[0019] Specifically, a target reshaping component is used to reshape a deep neural network to obtain an intelligent analysis body, wherein the reshaping of the deep neural network to obtain an intelligent analysis body includes: performing training operations on the deep neural network to obtain a deep neural network after each training operation is completed and used as an intelligent analysis body, including: using a CPLD device to implement the target reshaping component, which is used to reshape the deep neural network to obtain an intelligent analysis body, wherein the reshaping of the deep neural network to obtain an intelligent analysis body includes: performing training operations on the deep neural network to obtain a deep neural network after each training operation is completed and used as an intelligent analysis body;

[0020] A multi-stage processing mechanism, connected to the rear-end monitoring component and the target reshaping component respectively and comprising a primary processing component, a secondary processing component and a final processing component, wherein the primary processing component is used to extract the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the secondary processing component is used to obtain the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle, and the final processing component is connected to the secondary processing component and the primary processing component respectively, and is used to intelligently identify whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle;

[0021] An information transmission mechanism, connected to the intelligent analysis device, for transmitting the received intelligent identification results to a remote big data server through a network connection channel;

[0022] Among them, intelligently identifying whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel of the scene image signal, as well as the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle includes: using an intelligent analysis body to intelligently identify whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel of the scene image signal, as well as the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle;

[0023] Among them, intelligently identifying whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component based on the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel of the scene image signal, as well as the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle also includes: in the intelligent analysis body, the number of times the deep neural network is trained is proportional to the number of pixels constituting the scene image signal output by the rear-end monitoring component.

[0024] Compared with the A embodiment, the visual monitoring image content identification device shown in the B embodiment of the present invention may also include:

[0025] An electric quantity measuring mechanism, connected to the target remodeling component, the primary processing component, the secondary processing component and the final processing component, respectively, for measuring the current electric quantity used by the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively;

[0026] Wherein, the power measurement mechanism is respectively connected with the target remodeling component, the primary processing component, the secondary processing component and the final processing component, and is used to respectively measure the current power usage of each of the target remodeling component, the primary processing component, the secondary processing component and the final processing component, including: the power measurement mechanism includes a plurality of power measurement units, which are respectively connected with the target remodeling component, the primary processing component, the secondary processing component and the final processing component, so as to respectively measure the current power usage of each of the target remodeling component, the primary processing component, the secondary processing component and the final processing component;

[0027] Wherein, the power measurement mechanism includes a plurality of power measurement units, which are used to be connected to the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively, so as to complete the measurement of the current power usage of each of the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively, including: the plurality of power measurement units are a plurality of power sensing circuits, which are used to be connected to the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively, so as to complete the measurement of the current power usage of each of the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively;

[0028] Wherein, the plurality of power measurement units are a plurality of power sensing circuits, which are used to be connected to the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively, so as to complete the respective measurement of the current power usage of the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively, including: the structures of the plurality of power sensing circuits are the same;

[0029] Among them, the multiple power measurement units are multiple power sensing circuits, which are used to be connected to the target reshaping component, the primary processing component, the secondary processing component and the final processing component respectively to complete the separate measurement of the current power usage of the target reshaping component, the primary processing component, the secondary processing component and the final processing component respectively, and also include: the multiple power sensing circuits have the same power measurement upper limit value and power measurement lower limit value.

[0030] Compared to the A embodiment, the visual surveillance image content identification device shown in the C embodiment of the present invention may also include:

[0031] A fan heat dissipation mechanism is arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component, and is used to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with the heat dissipation services they need respectively;

[0032] Wherein, the fan heat dissipation mechanism is arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component, and is used to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with the heat dissipation services required by each of them, including: the fan heat dissipation mechanism includes a plurality of fan heat dissipation units, which are respectively arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with the heat dissipation services required by each of them;

[0033] And wherein, the fan cooling mechanism includes a plurality of fan cooling units, which are respectively arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with their respective required cooling services, including: the rated output power of the plurality of fan cooling units is equal.

[0034] In addition, in the visual monitoring image content identification device, based on the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel of the scene image signal, as well as the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle, intelligently identifying whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component also includes: using a content conversion function to represent a numerical conversion relationship in which the number of times the deep neural network is trained is proportional to the number of pixels constituting the scene image signal output by the rear-end monitoring component.

[0035] The visual monitoring image content identification device of the present invention aims to solve the technical problem in the prior art that image processing or special effects processing performed on the rear monitoring screen of the electric vehicle may affect the identification and early warning of rear-end collisions and other safety accidents. By introducing a multi-level processing mechanism with a customized hardware architecture and a targeted artificial intelligence model, the device is used to perform intelligent identification of whether the on-site image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component, thereby avoiding misjudgment in the identification and early warning of rear-end collisions and other safety accidents.

[0036] It is to be understood that, although the present invention has been disclosed as a preferred embodiment, the above embodiment is not intended to limit the present invention. For any technician familiar with the art, without departing from the scope of the technical solution of the present invention, the technical content disclosed above can be used to make many possible changes and modifications to the technical solution of the present invention, or modified into equivalent embodiments of equivalent changes. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. A visual surveillance image content identification device, characterized in that: The device comprises: A rear-end monitoring component, fixed at the rear end of the electric vehicle, is used to perform visual monitoring of the rear-end scene of the running electric vehicle, and is used to acquire and output corresponding on-site image signals; A target reshaping component is used to reshape the deep neural network to obtain an intelligent analysis body, wherein the reshaping of the deep neural network to obtain the intelligent analysis body includes: performing training operations on the deep neural network to obtain the deep neural network after completing each training operation and serving as the intelligent analysis body; A multi-stage processing mechanism, connected to the rear-end monitoring component and the target reshaping component respectively and comprising a primary processing component, a secondary processing component and a final processing component, wherein the primary processing component is used to extract the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the secondary processing component is used to obtain the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle, and the final processing component is connected to the secondary processing component and the primary processing component respectively, and is used to use an intelligent analysis body to intelligently identify whether the scene image signal is a BAYER image directly output by the photoelectric sensor of the rear-end monitoring component according to the contrast, signal-to-noise ratio, maximum noise amplitude, number of noise types and brightness values ​​corresponding to each pixel point of the scene image signal, the width of the driving road and the nearest car distance of the rear-end scene of the electric vehicle; Wherein, in the intelligent analysis body, the number of times the deep neural network is trained is proportional to the number of pixels constituting the on-site image signal output by the back-end monitoring component; Wherein, a content conversion function is used to represent a numerical conversion relationship in which the number of times the deep neural network is trained is proportional to the number of pixels constituting the on-site image signal output by the back-end monitoring component; An information transmission mechanism, connected to the intelligent analysis device, for transmitting the received intelligent identification results to a remote big data server through a network connection channel; An electric quantity measuring mechanism, connected to the target remodeling component, the primary processing component, the secondary processing component and the final processing component, respectively, for measuring the current electric quantity used by the target remodeling component, the primary processing component, the secondary processing component and the final processing component respectively; Wherein, the power measurement mechanism includes multiple power measurement units, which are used to connect with the target reshaping component, the primary processing component, the secondary processing component and the final processing component respectively, so as to complete the respective measurement of the current power usage of the target reshaping component, the primary processing component, the secondary processing component and the final processing component.

2. The visual surveillance image content identification device according to claim 1, characterized in that: The plurality of electric quantity measuring units are a plurality of electric quantity sensing circuits.

3. The visual surveillance image content identification device according to claim 2, characterized in that: The structures of the multiple power sensing circuits are the same.

4. The visual surveillance image content identification device according to claim 3, characterized in that: The plurality of power sensing circuits have the same power measurement upper limit value and power measurement lower limit value.

5. The visual surveillance image content identification device according to claim 1, characterized in that: The device also includes: A fan heat dissipation mechanism is arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component, and is used to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with the required heat dissipation services respectively.

6. The visual surveillance image content identification device according to claim 5, characterized in that: The fan cooling mechanism includes a plurality of fan cooling units, which are respectively arranged near the target reshaping component, the primary processing component, the secondary processing component and the final processing component to provide the target reshaping component, the primary processing component, the secondary processing component and the final processing component with the required cooling services respectively.

7. The visual surveillance image content identification device according to claim 6, characterized in that: The rated output powers of the multiple fan cooling units are equal.