WO2018113216A1

WO2018113216A1 - Dlp system based on machine learning

Info

Publication number: WO2018113216A1
Application number: PCT/CN2017/088987
Authority: WO
Inventors: 陈豪坤
Original assignee: 威创集团股份有限公司
Priority date: 2016-12-19
Filing date: 2017-06-19
Publication date: 2018-06-28
Also published as: CN106791487A

Abstract

Disclosed a DLP system based on machine learning, for use in resolving the technical problem of great difficulty in management of a DLP display system due to that the existing DLP display wall is a very huge system and uses a large number of chips and sensor devices to support the operation of the entire system together, and therefore, the development and maintenance difficulty can be imagined. According to embodiments of the present invention, the system comprises: a signal collection terminal, a network exchange terminal, a display terminal, and a machine learning module. The signal collection terminal, the display terminal, and the machine learning module are communicationally connected to the network exchange terminal. The machine learning module obtains all data streams of the signal collection terminal and the display terminal by means of the network exchange terminal, and the machine learning module sets corresponding data tags according to preset thresholds of monitoring index parameters to form a multilayer learning network.

Description

A DLP system based on machine learning

The present application claims priority to Chinese Patent Application No. 201611176925.7, entitled "A DLP System Based on Machine Learning", filed on December 19, 2016, the entire contents of which is incorporated herein by reference. in.

Technical field

The present invention relates to the field of DLP system technologies, and in particular, to a DLP system based on machine learning.

Background technique

DLP is the abbreviation of “Digital Light Processing”, which means digital light processing, which means that the technology first processes the image signal digitally and then projects the light. It is based on the digital micromirror component developed by TI (Texas Instruments, USA) - DMD (Digital Micromirror Device) to complete the display of visual digital information. To be specific, DLP projection technology uses digital micromirror wafers (DMDs) as the primary critical processing component to implement digital optical processing.

The principle is to pass the cold light source emitted by the UHP bulb through the condensing lens, homogenize the light through the Rod (light rod), and the processed light passes through a Color Wheel to separate the light into RGB three colors (or RGBW). Some more colors), some manufacturers use BSV LCD splicing technology to filter light transmission, then cast the color on the DMD chip by the lens, and finally reflect the projection on the projection screen through the projection lens.

DLP display wall is a very large system, which uses a large number of chips and sensor devices to support the operation of the whole system. Its development and later maintenance difficulty can be imagined, which makes the management of DLP display system very difficult. technical problem.

Summary of the invention

A DLP system based on machine learning provided by an embodiment of the present invention solves the current DLP The display wall is a very large system, which uses a large number of chips and sensor devices to support the operation of the whole system. The development and the difficulty of maintenance in the later stage can be imagined, resulting in the management of DLP display system is very difficult. problem.

A DLP system based on machine learning provided by an embodiment of the present invention includes:

Signal acquisition terminal, network switching terminal, display terminal and machine learning module;

The signal collecting end, the display end and the machine learning module are in communication connection with the network switching end;

The machine learning module acquires all data flows of the signal collection end and the display end by using the network switching end, and the machine learning module sets the data flow according to a preset threshold value of the monitoring indicator parameter. Corresponding data tags form a multi-layer learning network that includes key performance parameters to be monitored.

Preferably, the signal collecting end is at least two.

Preferably, the signal collecting end comprises:

Signal source video acquisition module, video front end processing module, IP coding module and network transmission module;

The signal source video collection module, the video front end processing module, the IP encoding module, and the network sending module are sequentially connected in communication.

Preferably, the display ends are at least two.

Preferably, the display end comprises:

a network receiving module, an IP decoding module, a video signal processing module, and a movement driving display module;

The network receiving module, the IP decoding module, the video signal processing module, and the movement driving display module are sequentially connected in communication.

Preferably, the machine learning based DLP system further comprises:

The signal scheduling platform is in communication with the network switching end.

Preferably, the machine learning based DLP system further comprises:

a network storage module, communicatively coupled to the network switch.

Preferably, the machine learning based DLP system further comprises:

A PC transcoding cluster is in communication with the network switching end.

Preferably, the machine learning based DLP system further comprises:

The optical machine is communicatively connected to the display end.

Preferably, the multi-layer learning network is a neural network.

It can be seen from the above technical solutions that the embodiments of the present invention have the following advantages:

A DLP system based on machine learning includes: a signal acquisition end, a network exchange end, a display end, and a machine learning module; a signal acquisition end, a display end, and a machine learning module are in communication connection with the network exchange end; The machine learning module acquires all data streams of the signal collecting end and the display end through the network switching end, and the machine learning module sets the data label corresponding to the data stream according to the preset threshold value of the monitoring index parameter to form a multi-layer learning network, and the data stream Includes key performance parameters to be monitored. In this embodiment, all the data streams of the signal collecting end and the display end are acquired by the machine learning module through the network switching end, and the machine learning module sets the data label corresponding to the data stream according to the preset threshold value of the monitoring index parameter to form a plurality of layers. Learning the network, the data flow includes the key performance parameters to be monitored, and the current DLP display wall is a very large system, which uses a large number of chips and sensor devices to jointly support the operation of the entire system, its development and later Maintenance difficulty can be imagined, resulting in technical problems that are very difficult to manage DLP display systems.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any inventive labor.

1 is a schematic structural diagram of an embodiment of a DLP system based on machine learning according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of an application example of FIG. 1. FIG.

detailed description

The DLP system based on machine learning provided by the embodiment of the invention solves the current DLP display wall is a very large system, which uses a large number of chips and sensor devices to jointly support the operation of the entire system, and its development and The difficulty of maintenance in the later stage can be imagined, resulting in technical problems that are very difficult to manage DLP display systems.

In order to make the object, the features and the advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

Referring to FIG. 1, an embodiment of a machine learning-based DLP system provided by an embodiment of the present invention includes:

Signal collecting end 1, network switching end 2, display end 3 and machine learning module 4;

The signal collecting end 1, the display end 2 and the machine learning module 3 are communicatively connected to the network switching end 4;

The machine learning module 4 acquires all the data streams of the signal collecting end 1 and the display end 3 through the network switching terminal 2, and the machine learning module 4 sets the threshold according to the threshold value of the preset monitoring index parameter. The data label corresponding to the data stream forms a multi-layer learning network. The data stream includes key performance parameters to be monitored (such as the temperature of the display end optical machine, the network speed of the switching network, and other parameters of the important devices that affect the overall performance of the system).

Further, the signal collecting end 1 is at least two.

Further, the signal collecting end 1 includes:

Further, the display terminals 3 are at least two.

Further, the display terminal 3 includes:

Further, the machine learning based DLP system further includes:

The signal scheduling platform 5 is in communication connection with the network switching end.

Further, the machine learning based DLP system further includes:

The network storage module 6 is in communication connection with the network switching end.

Further, the machine learning based DLP system further includes:

The PC transcoding cluster 7 is in communication connection with the network switching end.

Further, the machine learning based DLP system further includes:

The optical machine 8 is communicatively coupled to the display terminal.

Further, the multi-layer learning network is a neural network.

The following describes a specific application scenario. As shown in Figure 2, the application examples include:

Add the machine learning function module based on the overall architecture of the system, see the yellow background module of Figure 2 in the drawing. First, the function module can collect and monitor the parameters of the important devices that affect the overall performance of the system in real time. When these parameters are abnormal, the records are initially labeled, the log file is recorded, and the first layer learning network is formed, and then continues to be based on The learning mark of the first layer of the network as input, the construction of the second layer of "learning network", and so on, will help to understand the deeper factors of the system's abnormality.

Assume that “8186 chip temperature”, “bulb brightness” and “network data transmission speed” are respectively: T, L, and V dimensions are the monitoring objects of the system. The threshold ranges of allowable fluctuation are respectively: Td, Ld, Vd, corresponding tags are Tag_t, Tag_l, and Tag_v (the "1" flag is normal, and the "0" flag is abnormal).

The data stream is aggregated to the "machine learning" module through the network. The module will label the corresponding data according to the threshold setting of the previous monitoring indicator parameters, so the obtained "data element" is in this format: [T+-Td, L+- Ld, V+-Vd, Tag_t, Tag_l, Tag_v], and the number of layers of the network can correspond to the number of tags. The final output of the network is that the system is running normally or there is a problem. When there is a problem, it can be backtracked. The key device is where the most likely problem is.

In this simple neural network, when there is a problem in the system, we can query the label of the current "data element". If it is: [T+-Td, L+-Ld, V+-Vd, 0, 0, 1], then The reason for confirming the problem in the system is caused by the abnormality of the 8186 chip temperature (T) and the bulb brightness (L). It can also be traced back in the past: Check whether Tag_t and Tag_l have a causal relationship with each other and cause an abnormality. This will be a good way to confirm the deeper causes of the problem.

In this embodiment, the signal acquisition terminal is obtained through the network switching terminal through the machine learning module. All the data streams of the display end, the machine learning module sets the corresponding data label according to the preset threshold value of the monitoring index parameter, and forms a multi-layer learning network, which solves the current DLP display wall is a very large system, which uses a large number of The chip and the sensor device jointly support the operation of the entire system, and its development and later maintenance difficulty can be imagined, resulting in a very difficult technical problem in the management of the DLP display system. It can minimize the development cost; it can optimize the overall performance of the whole system, help to improve the adaptive ability of the system; and deepen the value that system testing can bring to the product.

A person skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the system, the device and the unit described above can refer to the corresponding process in the foregoing method embodiment, and details are not described herein again.

In the several embodiments provided by the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of the unit is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention contributes in essence or to the prior art or the technical solution All or part may be embodied in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform the present invention. All or part of the steps of the method described in the examples. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like. .

The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that The technical solutions described in the embodiments are modified, or the equivalents of the technical features are replaced by the equivalents of the technical solutions of the embodiments of the present invention.

Claims

A machine learning based DLP system, comprising:

Signal acquisition terminal, network switching terminal, display terminal and machine learning module;

The signal collecting end, the display end and the machine learning module are in communication connection with the network switching end;

The machine learning module acquires all data flows of the signal collection end and the display end by using the network switching end, and the machine learning module sets the data flow according to a preset threshold value of the monitoring indicator parameter. Corresponding data tags form a multi-layer learning network that includes key performance parameters to be monitored.
The machine learning-based DLP system according to claim 1, wherein the signal acquisition end is at least two.
The machine learning-based DLP system according to claim 2, wherein the signal acquisition end comprises:

Signal source video acquisition module, video front end processing module, IP coding module and network transmission module;

The signal source video collection module, the video front end processing module, the IP encoding module, and the network sending module are sequentially connected in communication.
The machine learning based DLP system according to claim 3, wherein the display end is at least two.
The machine learning-based DLP system according to claim 4, wherein the display end comprises:

a network receiving module, an IP decoding module, a video signal processing module, and a movement driving display module;

The network receiving module, the IP decoding module, the video signal processing module, and the movement driving display module are sequentially connected in communication.
The machine learning based DLP system according to claim 5, wherein the machine learning based DLP system further comprises:

The signal scheduling platform is in communication with the network switching end.
The machine learning based DLP system according to claim 6, wherein the machine learning based DLP system further comprises:

a network storage module, communicatively coupled to the network switch.
The machine learning based DLP system according to claim 7, wherein the machine learning based DLP system further comprises:

A PC transcoding cluster is in communication with the network switching end.
The machine learning based DLP system according to claim 8, wherein the machine learning based DLP system further comprises:

The optical machine is communicatively connected to the display end.
The machine learning based DLP system according to claim 1 or 9, wherein the multi-layer learning network is a neural network.