CN-115223094-B - Intelligent management monitoring system for production workshop based on ZYNQ

CN115223094BCN 115223094 BCN115223094 BCN 115223094BCN-115223094-B

Abstract

The invention relates to a ZYNQ-based intelligent management and monitoring system for a production workshop, and belongs to the field of artificial intelligence and embedding. The system comprises an image acquisition module, a ZYNQ system board and terminal equipment which are sequentially connected, wherein the image acquisition module is used for acquiring an image of a production workshop in real time, the ZYNQ system board is used for acquiring a ZYNQ executable target detection file, detecting the image of the production workshop by using the ZYNQ executable target detection file and outputting a detection result, whether workers in the production workshop wear a dustproof headgear, a mask and a working uniform or not and whether flames exist in the production workshop or not is detected, and the terminal equipment is used for displaying the detection result and sending an alarm according to the detection result. According to the invention, the ZYNQ system board is used for replacing a PC, so that the production cost is reduced, the system volume is reduced, the ZYNQ-based intelligent management monitoring system for the production workshop can be deployed at any place at any time, and the application range is enlarged.

Inventors

ZHANG XUEJUN
SU WENJUN
XU HUIBIN
LI BIN
LIU KAI

Assignees

广西大学
广西大学
凭祥市三诺数字科技有限公司
凭祥市三诺数字科技有限公司

Dates

Publication Date: 20260421
Application Date: 20220719
Priority Date: 20220719

Claims (7)

1. A ZYNQ-based intelligent management monitoring system for a production workshop is characterized by comprising an image acquisition module, a ZYNQ system board and terminal equipment which are connected in sequence; The image acquisition module is used for acquiring images of a production workshop in real time; The ZYNQ system board is used for acquiring a ZYNQ executable target detection file, detecting the production workshop image by using the ZYNQ executable target detection file and outputting a detection result, wherein the detection result is whether workers in the production workshop wear dust-proof hoods, masks and work uniforms or not and whether flames exist in the production workshop or not; the ZYNQ executable target detection file comprises the structural information of a trained YOLOv4-tiny network and the weight parameters of a trained YOLOv-tiny network, and the ZYNQ executable target detection file is obtained through an upper computer and specifically comprises the following steps: Preparing a training data set, wherein the training data set comprises an original image of a production workshop and an annotated image of the production workshop; training the YOLOv-tiny network by using the training data set to obtain a trained YOLOv-tiny network; Saving the weight parameters and the structure information of the trained YOLOv-tini network as a floating point model file; Quantizing the floating point model file to obtain a fixed point model file; Converting the fixed-point model file into the ZYNQ executable target detection file; the YOLOv-tiny network comprises a backbone feature extraction network, a feature pyramid and a YOLOhead output layer which are connected in sequence; The backbone feature extraction network is used for extracting features of the workshop image and comprises an input layer, a first CBL convolution layer, a second CBL convolution layer, a first Resblock convolution layer, a second Resblock convolution layer, a third Resblock convolution layer and a third CBL convolution layer which are sequentially connected; The feature pyramid is used for strengthening the features to obtain the strengthened features, and comprises a first convolution layer, a second convolution layer, an up-sampling layer and a connecting layer, wherein the first convolution layer is connected with the third CBL convolution layer; The YOLOhead output layer is used for detecting the reinforced features to obtain a detection result; the terminal equipment is used for displaying the detection result and sending out an alarm according to the detection result.
2. The ZYNQ-based production plant intelligent management monitoring system of claim 1, wherein the ZYNQ system board comprises an SD card for storing a ZYNQ acceleration system and the ZYNQ executable target detection file; in the SD card, the ZYNQ acceleration system calls the ZYNQ executable target detection file, and the ZYNQ acceleration system is also used for accelerating the detection process.
3. The ZYNQ-based production workshop intelligent management monitoring system according to claim 2, wherein the ZYNQ system board further comprises a processing system end, a programmable logic end, a USB interface, an Ethernet interface, a display screen interface and a serial port, wherein the processing system end is connected with the programmable logic end, and the processing system end is also connected with the USB interface, the Ethernet interface, the display screen interface and the serial port through an AXI communication bus respectively; the processing system end is a main control end of the ZYNQ system board and is used for providing clocks for the programmable logic end, the USB interface, the Ethernet interface, the display screen interface and the serial port; The programmable logic end is used for controlling the USB interface, the Ethernet interface, the display screen interface and the serial port; the USB interface is used for connecting the image acquisition module; The Ethernet interface is used for connecting an upper computer; the display screen interface is used for connecting a display screen of the terminal equipment; The serial port is used for serial port communication.
4. The ZYNQ-based production plant intelligent management monitoring system of claim 3, wherein the terminal device comprises a base, a display screen and an alarm; The display screen is connected with the alarm; the display screen and the alarm are arranged on the base; The display screen is used for displaying the detection result, and the alarm is used for sending out an alarm according to the detection result.
5. The ZYNQ-based production plant intelligent management monitoring system of claim 4, wherein the ZYNQ system board is disposed inside the base.
6. The ZYNQ-based production plant intelligent management monitoring system of claim 3, wherein the ZYNQ system board further comprises an on/off button.
7. The ZYNQ-based production plant intelligent management monitoring system of claim 1, wherein the ZYNQ-executable object detection file further comprises an instruction set of a DPU parallel acceleration YOLOv-tini network and a main program of ZYNQ operation.

Description

Intelligent management monitoring system for production workshop based on ZYNQ Technical Field The invention relates to the field of artificial intelligence and embedding, in particular to a ZYNQ-based intelligent management and monitoring system for a production workshop. Background With the great development of productivity of Chinese, accidents in production workshops are increasing, and people pay more attention to sanitation and safety of production workshops. Wherein, many production workshop accidents can be avoided, for example, when a dustproof headgear is worn, the hair can be prevented from falling onto the surfaces of parts or production machines, and the problems of unqualified product quality, damaged appliances and the like are caused; the working uniform detection can prevent accidents and losses caused by the random access of non-working personnel to the production workshop, and the flame detection can discover flame at the first time and take action in time to reduce the risk of fire spreading. Therefore, intelligent management and monitoring of the production plant becomes more important and even becomes an indispensable tool for the production plant. However, the existing intelligent management monitoring system of the production plant is realized on PC (personal computer), which is placed in a fixed monitoring room. In order to meet the real-time monitoring requirements of the production plant sites, the server side needs to deploy a large number of high-performance PCs to run high-complexity target detection algorithms and process image data from a plurality of production plant sites in real time. Although the PC has very excellent detection precision and speed, the power consumption of the equipment is increased, and the problems of higher requirements on the server, high cost and the like are not suitable for on-site deployment of a large-scale production workshop. Moreover, the production workshop site safety monitoring system realized on the PC has the defects of poor flexibility, large volume and the like, and is not suitable for being deployed to a temporary working site at any time. Disclosure of Invention The invention aims to provide a ZYNQ-based intelligent management and monitoring system for a production workshop, which aims to solve the problems of high cost and small application range of the intelligent management and monitoring system for the production workshop in the prior art. In order to achieve the above object, the present invention provides the following solutions: a ZYNQ-based intelligent management monitoring system for a production workshop comprises an image acquisition module, a ZYNQ system board and terminal equipment which are connected in sequence; The image acquisition module is used for acquiring images of a production workshop in real time; The ZYNQ system board is used for acquiring a ZYNQ executable target detection file, detecting the production workshop image by using the ZYNQ executable target detection file and outputting a detection result, wherein the detection result is whether workers in the production workshop wear dust-proof hoods, masks and work uniforms or not and whether flames exist in the production workshop or not; the terminal equipment is used for displaying the detection result and sending out an alarm according to the detection result. Optionally, the ZYNQ system board includes an SD card for storing a ZYNQ acceleration system and the ZYNQ executable target detection file; in the SD card, the ZYNQ acceleration system calls the ZYNQ executable target detection file, and the ZYNQ acceleration system is also used for accelerating the detection process. Optionally, the ZYNQ system board further comprises a processing system end, a programmable logic end, a USB interface, an Ethernet interface, a display screen interface and a serial port, wherein the processing system end is connected with the programmable logic end and is also connected with the USB interface, the Ethernet interface, the display screen interface and the serial port through an AXI communication bus respectively; the processing system end is a main control end of the ZYNQ system board and is used for providing clocks for the programmable logic end, the USB interface, the Ethernet interface, the display screen interface and the serial port; The programmable logic end is used for controlling the USB interface, the Ethernet interface, the display screen interface and the serial port; the USB interface is used for connecting the image acquisition module; The Ethernet interface is used for connecting an upper computer; the display screen interface is used for connecting a display screen of the terminal equipment; The serial port is used for serial port communication. Optionally, the terminal device comprises a base, a display screen and an alarm; The display screen is connected with the alarm; the display screen and the alarm are arranged on the base; The display screen is u