Image processing type intelligent identification system

Abstract

The invention relates to an image processing type intelligent identification system comprising a model storage mechanism used for completing model storage of an intelligent identification model by storing various model parameters of the intelligent identification model; an intelligent identification mechanism, the method is used for adopting an intelligent identification model based on various camera parameters of a spherical camera device. And intelligently identifying the image processing type of the Bayer image mapped into the wide-area imaging picture according to each gray value corresponding to each pixel point of the wide-area imaging picture and each gray value corresponding to the Bayer image corresponding to the wide-area imaging picture. According to the method and the device, the image processing type of the Bayer image mapped into the wide-area imaging picture can be intelligently identified based on various visual association information by adopting the intelligent identification model which is designed in a targeted manner, so that the image processing type of the Bayer image is judged according to the content of the processed image, and the retrograde judgment process of image processing is completed.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to an image processing type intelligent recognition system. Background Art

[0002] Common types of image processing include image enhancement, image restoration, image filtering, image sharpening, image interpolation, and image distortion correction. For example, image enhancement and restoration for the purpose of improving image quality are widely used for some difficult-to-obtain images or images obtained under very poor shooting conditions. For example, photos of the earth or other planets taken from space, biomedical images taken with electron microscopes or X-rays, etc. The goal of image enhancement is to improve the quality of the image, such as increasing contrast, removing blur and noise, correcting geometric distortion, etc.; image restoration is a technique that attempts to estimate the original image when assuming a known blur or noise model.

[0003] However, currently it is possible to determine the type of image processing that needs to be performed on an image to be processed based on its content. Conversely, it is difficult to determine the type of image processing that has been processed based on its content, which leads to considerable obstacles to the subsequent judgment of image content and the restoration of image content. Summary of the invention

[0004] In order to solve the technical problems in the related fields, the present invention provides an image processing type intelligent recognition system, which obtains various camera parameters of a spherical camera device, and simultaneously obtains various grayscale values ​​corresponding to various pixel points of a wide-area imaging picture and various grayscale values ​​corresponding to a Bayer image corresponding to the wide-area imaging picture, wherein the Bayer image corresponding to the wide-area imaging picture is an image signal directly output by an image sensing component of the spherical camera device, which has not been image processed and corresponds to the wide-area imaging picture, thereby providing a plurality of basic data for intelligent recognition of the image processing type that the Bayer image has undergone in order to be mapped into the wide-area imaging picture, and then adopts a customized structure intelligent recognition model to intelligently recognize the image processing type that the Bayer image has undergone in order to be mapped into the wide-area imaging picture, wherein the structure of the intelligent recognition model is customized in that a feedforward neural network that has undergone multiple trainings is output as an intelligent recognition model, and the number of trainings performed on the feedforward neural network is proportional to the wide-area viewing angle corresponding to the wide imaging lens of the spherical camera device, and specifically, the various camera parameters of the spherical camera device are the minimum illumination, horizontal resolution and CCD size of the spherical camera device, thereby being able to judge the image processing type that has undergone the processed image according to its content.

[0005] According to the present invention, there is provided an image processing type intelligent recognition system, the system comprising: A multiple training device, used to output a feedforward neural network that has been trained multiple times as an intelligent recognition model, wherein the feedforward neural network is trained multiple times to obtain a feedforward neural network that has been trained multiple times; A spherical camera device is used to perform camera processing on the field of view of a wide imaging lens of the spherical camera device to obtain and output a corresponding wide-area imaging picture; A model storage mechanism, connected to the multiple training device, for completing the model storage of the intelligent recognition model by storing various model parameters of the intelligent recognition model; an intelligent recognition mechanism, connected to the multiple training device, the spherical camera device and the model storage mechanism, respectively, for obtaining various camera parameters of the spherical camera device, and simultaneously obtaining various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture, and using an intelligent recognition model to intelligently recognize the type of image processing that the Bayer image has undergone for mapping into the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen are obtained, and an intelligent recognition model is used to intelligently identify the image processing type that the Bayer image has undergone for mapping into the wide-area imaging screen based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen. The camera parameters of the spherical camera device are the minimum illumination, horizontal resolution and CCD size of the spherical camera device; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture are obtained, and an intelligent recognition model is used to intelligently identify the type of image processing that the Bayer image has undergone for mapping to the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture. The Bayer image corresponding to the wide-area imaging picture is an image signal directly output by the image sensing component of the spherical camera device, which has not been image processed and corresponds to the wide-area imaging picture; Among them, the multiple training device is used to output the feedforward neural network that has been trained multiple times as an intelligent recognition model, wherein the multiple training of the feedforward neural network to obtain the feedforward neural network that has been trained multiple times includes: the number of trainings performed on the feedforward neural network is proportional to the wide-range viewing angle corresponding to the wide imaging lens of the spherical camera device.

[0006] It can be seen that the present invention has at least the following three important inventive concepts: First, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixel points of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture are obtained, wherein the Bayer image corresponding to the wide-area imaging picture is an image signal directly output by the image sensing component of the spherical camera device, which has not been image processed and corresponds to the wide-area imaging picture, thereby providing multiple basic data for intelligent identification of the image processing type that the Bayer image has undergone in order to be mapped into the wide-area imaging picture; Second, a customized intelligent recognition model is used to intelligently recognize the type of image processing that the Bayer image undergoes when mapped into a wide-area imaging picture. The structure of the intelligent recognition model is customized in that a feedforward neural network that has been trained multiple times is output as the intelligent recognition model, and the number of times the feedforward neural network is trained is proportional to the wide-area viewing angle corresponding to the wide imaging lens of the spherical camera device. Third: Specifically, the various imaging parameters of the spherical imaging device are the minimum illumination, horizontal resolution and CCD size of the spherical imaging device, thereby providing reliable basic information for intelligent recognition processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein: Figure 1 FIG. 1 is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a first embodiment of the present invention.

[0008] Figure 2 FIG. 1 is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a second embodiment of the present invention.

[0009] Figure 3 FIG. 4 is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a third embodiment of the present invention. DETAILED DESCRIPTION

[0010] The embodiment of the image processing type intelligent recognition system of the present invention will be described in detail below with reference to the accompanying drawings.

[0011] Figure 1This is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a first embodiment of the present invention, wherein the system comprises: A multiple training device, used to output a feedforward neural network that has been trained multiple times as an intelligent recognition model, wherein the feedforward neural network is trained multiple times to obtain a feedforward neural network that has been trained multiple times; Specifically, a SOC device may be selected to implement the multiple training device, which is used to output the feedforward neural network after multiple training as an intelligent recognition model, wherein the feedforward neural network is trained multiple times to obtain the feedforward neural network after multiple training; A spherical camera device is used to perform camera processing on the field of view of a wide imaging lens of the spherical camera device to obtain and output a corresponding wide-area imaging picture; A model storage mechanism, connected to the multiple training device, for completing the model storage of the intelligent recognition model by storing various model parameters of the intelligent recognition model; an intelligent recognition mechanism, connected to the multiple training device, the spherical camera device and the model storage mechanism, respectively, for obtaining various camera parameters of the spherical camera device, and simultaneously obtaining various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture, and using an intelligent recognition model to intelligently recognize the type of image processing that the Bayer image has undergone for mapping into the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen are obtained, and an intelligent recognition model is used to intelligently identify the image processing type that the Bayer image has undergone for mapping into the wide-area imaging screen based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen. The camera parameters of the spherical camera device are the minimum illumination, horizontal resolution and CCD size of the spherical camera device; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture are obtained, and an intelligent recognition model is used to intelligently identify the type of image processing that the Bayer image has undergone for mapping to the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture. The Bayer image corresponding to the wide-area imaging picture is an image signal directly output by the image sensing component of the spherical camera device, which has not been image processed and corresponds to the wide-area imaging picture; The multiple training device is used to output the feedforward neural network after multiple training as an intelligent recognition model, wherein the multiple training of the feedforward neural network to obtain the feedforward neural network after multiple training includes: the number of trainings performed on the feedforward neural network is proportional to the wide-area viewing angle corresponding to the wide imaging lens of the spherical camera device; Among them, the spherical camera device is used to perform camera processing on the field of view of the wide imaging lens of the spherical camera device to obtain and output the corresponding wide-area imaging picture, including: the wide-area viewing angle corresponding to the wide imaging lens of the spherical camera device is greater than or equal to 270 degrees.

[0012] Figure 2 FIG. 1 is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a second embodiment of the present invention.

[0013] and Figure 1 different, Figure 2 The image processing type in the intelligent recognition system can also include the following components: A data monitoring device, comprising a plurality of monitoring cameras facing the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, for monitoring the real-time status of the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism in a time-sharing manner; Wherein, the data monitoring device comprises a plurality of monitoring cameras facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, including: the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism are respectively connected to the data monitoring device using different CSI interfaces; Wherein, the data monitoring device includes a plurality of monitoring cameras facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism. The data monitoring device also includes: the plurality of monitoring cameras facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively have the same structure; Wherein, the data monitoring device includes a plurality of monitoring cameras facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism. It also includes: a plurality of monitoring cameras with built-in image sensing units facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively; Among them, a plurality of monitoring cameras respectively facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism have built-in image sensing units: the image sensing units are CMOS sensing units or CCD sensing units.

[0014] Figure 3 FIG. 4 is a schematic diagram of the internal structure of an image processing type intelligent recognition system according to a third embodiment of the present invention.

[0015] and Figure 1 different, Figure 3 The image processing type in the intelligent recognition system can also include the following components: A content display device is connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, and is used to simultaneously display various status information of the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and further includes: the content display device is a liquid crystal display screen or an LED display array; Wherein, the quartz oscillator device is respectively connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used to simultaneously provide the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism with the working clock pulses they need; And wherein, the quartz oscillator device is respectively connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used to simultaneously provide the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism with the working clock pulses they need respectively, including: the working clock pulses they need respectively provided to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism simultaneously are all square wave waveforms.

[0016] In addition, in the image processing type intelligent recognition system, an intelligent recognition mechanism is respectively connected to the multiple training device, the spherical camera device and the model storage mechanism, and is used to obtain various camera parameters of the spherical camera device, and simultaneously obtain various grayscale values ​​corresponding to various pixel points of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture. An intelligent recognition model is used to intelligently identify the image processing type that the Bayer image has undergone for mapping to the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixel points of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture. The intelligent recognition mechanism includes: a first acquisition component, a second acquisition component, a synchronous processing component and an identification execution component are built in the intelligent recognition mechanism, and the synchronous processing component is used to realize the synchronous control of the respective actions of the first acquisition component, the second acquisition component and the identification execution component.

[0017] The image processing type intelligent recognition system of the present invention solves the technical problem in the prior art that it is difficult to judge the type of image processing that has been processed based on the content of the processed image. By adopting a targeted intelligent recognition model, the image processing type that the Bayer image has undergone as mapped into a wide-area imaging screen can be intelligently identified based on various visual association information. The image processing type that has been processed can be judged based on the content of the processed image, thereby completing the reverse judgment process of image processing and solving the above technical problem.

[0018] Although the preferred embodiment of the present invention has been illustrated and described, it should be understood that the present invention is not limited thereto. Numerous modifications, changes, variations, substitutions and equivalents will occur to those skilled in the art without departing from the spirit and scope of the present invention as defined by the claims.

Claims

1. An image processing type intelligent recognition system, characterized in that: The system comprises: A multiple training device, used to output a feedforward neural network that has been trained multiple times as an intelligent recognition model, wherein the feedforward neural network is trained multiple times to obtain a feedforward neural network that has been trained multiple times; A spherical camera device is used to perform camera processing on the field of view of a wide imaging lens of the spherical camera device to obtain and output a corresponding wide-area imaging picture; A model storage mechanism, connected to the multiple training device, for completing the model storage of the intelligent recognition model by storing various model parameters of the intelligent recognition model; an intelligent recognition mechanism, connected to the multiple training device, the spherical camera device and the model storage mechanism, respectively, for obtaining various camera parameters of the spherical camera device, and simultaneously obtaining various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture, and using an intelligent recognition model to intelligently recognize the type of image processing that the Bayer image has undergone for mapping into the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen are obtained, and an intelligent recognition model is used to intelligently identify the image processing type that the Bayer image has undergone for mapping into the wide-area imaging screen based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging screen and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging screen. The camera parameters of the spherical camera device are the minimum illumination, horizontal resolution and CCD size of the spherical camera device; Among them, various camera parameters of the spherical camera device are obtained, and at the same time, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture are obtained, and an intelligent recognition model is used to intelligently identify the type of image processing that the Bayer image has undergone for mapping to the wide-area imaging picture based on various camera parameters of the spherical camera device, various grayscale values ​​corresponding to various pixels of the wide-area imaging picture and various grayscale values ​​corresponding to the Bayer image corresponding to the wide-area imaging picture. The Bayer image corresponding to the wide-area imaging picture is an image signal directly output by the image sensing component of the spherical camera device, which has not been image processed and corresponds to the wide-area imaging picture; Among them, the multiple training device is used to output the feedforward neural network that has been trained multiple times as an intelligent recognition model, wherein the multiple training of the feedforward neural network to obtain the feedforward neural network that has been trained multiple times includes: the number of trainings performed on the feedforward neural network is proportional to the wide-range viewing angle corresponding to the wide imaging lens of the spherical camera device.

2. The image processing type intelligent recognition system according to claim 1, characterized in that: The spherical camera device is used to process the field of view of the wide imaging lens of the spherical camera device to obtain and output the corresponding wide-area imaging picture, including: the wide-area viewing angle corresponding to the wide imaging lens of the spherical camera device is greater than or equal to 270 degrees.

3. The image processing type intelligent recognition system according to claim 2, characterized in that: The system further comprises: A data monitoring device, comprising a plurality of monitoring cameras facing the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, for monitoring the real-time status of the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism in a time-sharing manner; Among them, the data monitoring device includes multiple monitoring cameras facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism respectively, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism. The multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism are connected to the data monitoring device respectively using different CSI interfaces.

4. The image processing type intelligent recognition system according to claim 3, characterized in that: The data monitoring device includes a plurality of monitoring cameras respectively facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism. It also includes: the structures of the plurality of monitoring cameras respectively facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism are the same.

5. The image processing type intelligent recognition system according to claim 4, characterized in that: The data monitoring device includes a plurality of monitoring cameras respectively facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used for time-sharing monitoring of the real-time status of the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism. It also includes: a plurality of monitoring cameras respectively facing the multiple training devices, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and built-in image sensing units.

6. The image processing type intelligent recognition system according to claim 5, characterized in that: A plurality of monitoring cameras are provided with built-in image sensing units respectively facing the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism: the image sensing units are CMOS sensing units or CCD sensing units.

7. The image processing type intelligent recognition system according to claim 2, characterized in that: The system further comprises: The content display device is respectively connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used to simultaneously display various status information of the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and also includes: the content display device is a liquid crystal display screen or an LED display array.

8. The image processing type intelligent recognition system according to claim 7, characterized in that: A quartz oscillator device is respectively connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used to simultaneously provide the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism with the required working clock pulses.

9. The image processing type intelligent recognition system according to claim 8, characterized in that: A quartz oscillator device is respectively connected to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism, and is used to simultaneously provide the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism with the working clock pulses they need, including: the working clock pulses they need provided to the multiple training device, the spherical camera device, the model storage mechanism and the intelligent recognition mechanism are all square wave waveforms.

Images