CN213990849U

CN213990849U - Image detection system based on MEC edge cloud

Info

Publication number: CN213990849U
Application number: CN202022479962.3U
Authority: CN
Inventors: 郭文康; 温振山; 肖益珊
Original assignee: Yitong Century Internet Of Things Research Institute Guangzhou Co ltd
Current assignee: Guangdong Yitong Lianyun Intelligent Information Co.,Ltd.
Priority date: 2020-10-30
Filing date: 2020-10-30
Publication date: 2021-08-17
Anticipated expiration: 2030-10-30

Abstract

The utility model discloses an image detection system based on MEC edge cloud, include: the wireless transmission camera is used for acquiring image data and uploading the image data to the MEC edge cloud in a mobile network mode; the MEC edge cloud is used for processing the image data, sending a processing result to the Internet of things cloud platform and sending a first instruction to the wireless transmission camera; and the Internet of things cloud platform receives the processing result from the MEC edge cloud, displays the processing result through a visual interface, and sends a second instruction to the MEC edge cloud. The utility model discloses form a closed loop with this triplex of wireless transmission camera, MEC edge cloud and thing networking cloud platform, utilize the cloud computing power of MEC edge cloud, carry out data processing at MEC edge cloud, alleviate the pressure of inserting side and platform end at equipment for the bad outward appearance of product detects, improves the discernment rate of accuracy, but wide application in image detection technical field.

Description

Image detection system based on MEC edge cloud

Technical Field

The utility model relates to an image detection technical field especially relates to an image detection system based on MEC edge cloud.

Background

With the development of 5G networks, because 5G networks have the characteristics of ultra-high bandwidth, ultra-high density and ultra-low time delay, the load to be carried by mobile networks is dozens of times or hundreds of times before, and delay is not delayed, so that the pressure on the networks is increased more and more. And the MEC attaches computing power to the edge of the network, realizes the localization of the service, can effectively reduce service delay, bandwidth overhead and terminal cost, improves service experience and data safety, and provides effective support for novel service with human center and universal interconnection application with object as center.

The prior logo detection system is locally provided with a set of detection system, so that a set of computer hardware (calculator) is required to support in a deployment field, the deployment is relatively troublesome, the computer hardware has limited computing power, and the time consumption is long when complex images are processed.

Interpretation of terms:

MEC edge cloud: namely, multi-access Edge Computing (Mobile Edge Computing), provides an IT service environment and cloud Computing capability for the Edge of a Mobile network, and cancels the delay related to backhaul by performing partial caching, data transmission and computation at the Edge of the Mobile network, so that millisecond-level application can be finally realized.

The product logo detection system is used for identifying whether the logo of the product is correct or whether the logo has an offset position or not by locally deploying a set of product logo detection software.

SUMMERY OF THE UTILITY MODEL

In order to solve the technical problem, the utility model aims at providing an image detection system based on MEC edge cloud.

The utility model adopts the technical proposal that:

an image detection system based on MEC edge cloud, comprising:

the wireless transmission camera is used for acquiring image data and uploading the image data to the MEC edge cloud in a mobile network mode;

the MEC edge cloud is used for processing the image data, sending a processing result to the Internet of things cloud platform and sending a first instruction to the wireless transmission camera;

and the Internet of things cloud platform receives the processing result from the MEC edge cloud, displays the processing result through a visual interface, and sends a second instruction to the MEC edge cloud.

Further, the wireless transmission camera comprises a camera module, a wireless transmission module and a fixing frame, wherein the camera module is connected with the wireless transmission module in a serial port mode, and the fixing frame is used for fixing the camera module so that the focal length of the camera module can reach the best.

Further, the MEC edge cloud comprises an image recognition module, a neural network model is arranged on the image recognition module, and the neural network model comprises a first convolution layer, a first pooling layer, a second convolution layer, a second pooling layer, a full-connection layer and a full-connection output layer.

Further, the MEC edge cloud uploads the processing result to the Internet of things cloud platform through an MQTT communication protocol.

Further, a display screen is arranged on the Internet of things cloud platform and used for displaying the processing result.

The utility model has the advantages that: the utility model discloses form a closed loop with this triplex of wireless transmission camera, MEC edge cloud and thing networking cloud platform, utilize the cloud computing power of MEC edge cloud, carry out data processing at MEC edge cloud, alleviate the pressure of inserting side and platform end at equipment for the bad outward appearance of product detects, improves the discernment rate of accuracy.

Drawings

The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a topological diagram of an image detection system based on MEC edge cloud in an embodiment of the present invention.

Detailed Description

This section will describe in detail the embodiments of the present invention, preferred embodiments of the present invention are shown in the attached drawings, which are used to supplement the description of the text part of the specification with figures, so that one can intuitively and vividly understand each technical feature and the whole technical solution of the present invention, but they cannot be understood as the limitation of the protection scope of the present invention.

In the description of the present invention, it should be understood that the orientation or positional relationship indicated with respect to the orientation description, such as up, down, front, rear, left, right, etc., is based on the orientation or positional relationship shown in the drawings, and is only for convenience of description and simplification of description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention.

In the description of the present invention, a plurality of means are one or more, a plurality of means are two or more, and the terms greater than, less than, exceeding, etc. are understood as not including the number, and the terms greater than, less than, within, etc. are understood as including the number. If the first and second are described for the purpose of distinguishing technical features, they are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated.

In the description of the present invention, unless there is an explicit limitation, the words such as setting, installation, connection, etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in combination with the specific contents of the technical solution.

As shown in fig. 1, the present embodiment provides an image detection system based on MEC edge cloud, including:

In the embodiment, the internet of things cloud platform mainly has the functions of issuing collected instructions and receiving processing results from the MEC edge cloud, and the processing results are displayed on a visual interface. The MEC edge cloud provides strong cloud computing capability, and the processing result is returned to the Internet of things cloud platform after the image is processed by deploying a detection algorithm in the MEC edge cloud. The wireless transmission camera is bottom hardware for acquiring images, combines the camera with the wireless communication module, and uploads image data to the MEC edge cloud in a mobile network mode. The three parts form a closed loop for detecting the bad appearance of the product and improving the identification accuracy.

The wireless transmission camera can be divided into three parts: camera module, wireless transmission module, mount. The camera module is connected with the wireless transmission module through a serial port, and after receiving the acquisition instruction, the camera starts to acquire images and then transmits the images to the MEC edge cloud by using the wireless transmission module. The fixed mount is mainly used for fixing the camera on a production line, so that the focal length of the camera reaches the best and the definition of an image is ensured.

The algorithm deployed in the MEC edge cloud is divided into a preprocessing module, an image recognition module and a detection result returning module.

A preprocessing module: the image is preprocessed based on OpenCV. The preprocessing comprises the steps of obtaining a camera color image, converting the camera color image into a gray image, adopting a binarization algorithm and a dilation algorithm, extracting each outline, cutting the binarized image according to each outline to obtain each sub-image, and rotating the sub-image according to a rotating angle if the minimum circumscribed rectangle of the outline has the rotating angle. The binarization algorithm can change the image background into black and the image content with important information into white. Meanwhile, the abnormal contour is filtered according to the practical application, so that the number of the sub-images is reduced, and the calculation amount of a subsequent program is reduced. The sub-image is resized to become a 48x48 pixel thumbnail. The processed image can improve the effect under the condition of non-uniform light. And then, the method is used for finding out the outline, and rotating the image according to the rotation angle of the minimum circumscribed rectangle of the outline so as to enable the image to be positive. This allows the camera to be offset a small distance or rotated some degree. As long as the target to be recognized can be in the visual field range of the camera, the area to be recognized can be automatically extracted as a subimage, and the rotation correction direction is realized.

An image recognition module: is a LeNet neural network model based on Tensorflow. The model has been trained from the samples in advance. The function is to classify the sub-images and obtain the probability of each class. The network algorithm can quickly identify scenes with fewer categories through testing to obtain results. The following is a detailed description of this section:

LeNet is a neural network. Is one of the earliest convolutional neural networks. The network layer number is 6, and the initial application is in character recognition and has a relatively high recognition rate. Due to the rapid development of the semiconductor industry, with the increasing computing power, the network algorithm occupies relatively less computing resources.

The model structure is as follows: convolutional layer 1 (i.e., the first convolutional layer), pooling layer 1 (i.e., the first pooling layer), convolutional layer 2 (i.e., the second convolutional layer), pooling layer 2 (i.e., the second pooling layer), fully-connected layer, and fully-connected output layer.

The input image size was adjusted to 48 pixels long by 48 pixels wide by 3 channels. The number of convolution kernels of convolution layer 1 is 20, and each size is (5, 5). The activation function is a linear rectification function "ReLU". The pooling window size for subsequent pooling layer 1 is (2,2) and the downsampling factor is (2, 2). The number of convolution kernels of convolution layer 2 is 50, and each size is (5, 5). The activation function is a linear rectification function "ReLU". The pooling window size of pooling layer 2 is (2,2) and the downsampling factor is (2, 2). The number of neurons in the fully-connected layer was 500, and the neurons were arranged in a row.

And a detection result returning module: and returning the processing result to the cloud platform of the Internet of things, and finishing the identification of the image flaws after the processing of the first two parts. And finally, uploading the processing result to an Internet of things cloud platform by using an MQTT communication protocol, and carrying out visual interface display.

The Internet of things cloud platform can be divided into two parts, namely issuing of an acquisition instruction and displaying of a processing result. The first part is that an acquisition instruction is established on the Internet of things cloud platform, when an image needs to be acquired, the instruction is issued, and after the equipment receives the instruction, the camera is opened for acquisition. The second part is that the processing result returned by the MEC edge cloud is displayed, so that managers can more visually check the integrity of product detection.

In summary, compared with the prior art, the image detection system based on the MEC edge cloud of the embodiment has the following beneficial effects:

(1) only a wireless transmission camera needs to be deployed, and the deployment in a working site is simple. The hardware requirement of the terminal is reduced, and a computer is not required to be added locally.

(2) And the image processing speed can be improved by utilizing the cloud computing capability of the MEC edge cloud. And the cloud at the edge of the MEC is used for accessing and processing data, so that the pressure on the access side and the platform end of the equipment is relieved.

(3) The remote control acquisition function is carried out on the terminal equipment through the command issued by the cloud platform of the Internet of things, so that remote management and centralized management are realized.

(4) Visual interface viewing is provided on the cloud platform of the Internet of things, and meanwhile, warning reminding is provided so that managers can know the appearance of defective products at the first time.

While the preferred embodiments of the present invention have been described, the present invention is not limited to the above embodiments, and those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and such equivalent modifications or substitutions are intended to be included within the scope of the present invention as defined by the appended claims.

Claims

1. An image detection system based on MEC edge cloud, comprising:

2. The image detection system based on the MEC edge cloud as claimed in claim 1, wherein the wireless transmission camera comprises a camera module, a wireless transmission module and a fixing frame, the camera module is connected with the wireless transmission module by means of a serial port, and the fixing frame is used for fixing the camera module so as to optimize the focal length of the camera module.

3. The MEC edge cloud based image detection system of claim 1, wherein the MEC edge cloud uploads the processing results to the IOT cloud platform via an MQTT communication protocol.

4. The image detection system based on the MEC edge cloud of claim 1, wherein a display screen is arranged on the cloud platform of the Internet of things and used for displaying the processing result.