Intelligent anti-theft and fire warning system based on AVR32

1 Introduction

This project uses Atmel's AVR32 AT32UC3A processor as the core and uClinux operating system as the platform. It uses a wide-angle camera to collect video (or images) and uses algorithms to implement functions such as face recognition (indoor anti-theft) or flame recognition (outdoor fire prevention). On this basis, other functions are expanded:

(1) Through the Ethernet port and related protocols, the system platform can communicate with the remote computer, thereby realizing the transmission of multimedia data and remote monitoring functions.

(2) Through the SD module and the expanded GPRS (GSM) module, the system platform can communicate with the personal terminal device PDA, thereby realizing functions such as SMS or phone reminders, alarms and early warnings.

The hardware platform used in this project is EVK1100. Considering the chip processing power and real-time requirements, it is possible to use AT32AP7 series chips with stronger performance and better signal processing capabilities.

2. Project Background

The illegal crime of "burglary" is becoming increasingly rampant around the world and has become a global problem. The world invests nearly $20 billion every year to prevent burglary, while the losses caused by burglary are as high as more than $40 billion each year. However, the traditional means of human defense (manpower security measures, such as patrols, guards, etc.) and physical defense (such as anti-theft windows, anti-theft doors, safes, etc.) can no longer effectively prevent the continued rise in theft rate. In such a severe situation, a new and more effective security prevention method is about to emerge.

On the other hand, according to statistics, the world's land affected by forest and wildland fires is about 350 million hectares every year. These fires not only bring huge losses of life and property to humans, but also further aggravate problems such as climate change, air pollution and loss of biodiversity. Therefore, there is an urgent need for a measure to prevent the occurrence and spread of forest and wildland fires. The Food and Agriculture Organization of the United Nations once pointed out that the main cause of forest and wildland fires is often some human factors, and taking comprehensive measures such as prevention, monitoring, assessment and early warning can effectively prevent the occurrence of forest and wildland fires, thereby reducing the occurrence of fires and reducing the losses caused by fires.

This project develops an intelligent anti-theft and fire early warning system to prevent the above social problems. The system is particularly suitable for residential homes, supermarkets, banks, forest fire prevention and control and other places, and has a wide range of applications.

2. Demand Analysis

1. Functional requirements

The system consists of two parts: the data acquisition terminal and the receiving terminal. The receiving terminal can be a personal mobile terminal PDA or a remote monitoring PC. The communication between the system and the personal terminal is mainly based on the GSM module (sending text messages, etc.) or the GPRS module (surfing the Internet), and can even use the latest 3G technology. The communication between the system and the terminal PC needs to be achieved through the IP/TCP protocol, using the socket mechanism to achieve network communication and video data reception. Here, the system platform is actually regarded as a server, and the transmission channel can choose wireless or wired channels. The specific system architecture is shown in the figure below:

2. Performance requirements

Data collection: In order to collect good video data, the system should be placed in a well-lit place so that useful data can be collected. Of course, you can also use a photosensitive sensor to automatically turn on a small light when the system's working environment becomes dark.

Data processing: Since the application requires strong real-time performance, the system has low delay in data collection, processing and sending.

Considering the system's processing and storage capabilities, the camera's pixels should not be too high.

The processor should be able to port some commonly used microprocessor operating systems, such as uClinux.

The accuracy of the algorithm should be high enough to avoid false alarms and unnecessary trouble. The communication part with the police station or fire department can be connected externally as needed.

3. Solution Design

1. System function implementation principle

As shown in the figure below, the system is mainly composed of image or video acquisition, core processing unit, GSM module, Ethernet module, photosensitive sensor module, etc. The data collected by the camera is the input part of the system. AVR32 AT32UC3A and uClinux form the core processing unit of the intelligent anti-theft and fire prevention system. The camera captures the surrounding situation at all times and sends the signal to the AVR3 AT32UC3A processor. The system uClinux running on the AVR32 AT32UC3A is responsible for controlling the operation of the entire system and the implementation of the core algorithm.

2. Hardware platform selection and resource allocation

The EVK1100 development board supports the AT32UC3A chip, and its peripheral devices include: an Ethernet port, light and temperature potentiometer sensors, 20 characters × 4 lines of blue LED lights, JTAG, Nexus, USART, USB, TWI and SPI interfaces, and an SD/MMC card reader. The main circuit module diagrams are:

Clock and reset module circuit:

SD/MMC module circuit:

SPI module circuit:

JTAG module circuit:

This project is based on the Atmel AT32UC3A EVK1100 development board and some corresponding expansion boards. It uses a camera to perform real-time monitoring and transmits data to a remote PC through the network. This system platform mainly includes:

Camera

Choose a wide-angle CCD camera to facilitate all-round monitoring. The pixel of the camera here should not be too high, otherwise it may lead to insufficient storage space and inability to receive data completely.

Network communication module

The EVK1100 development board has a built-in Ethernet interface that supports wired channel network transmission. If the system needs to be placed outdoors, a corresponding Wi-Fi module needs to be expanded. Through the Ethernet module, on the one hand, data can be forwarded from the system platform to the remote PC in a timely manner, and on the other hand, control signals from the PC can be received in a timely manner.

GPRS module (GSM module)

The GPRS (GSM) module is mainly used for communication between the system and the personal mobile terminal device PDA. Because we cannot monitor on the remote PC all the time. Therefore, when something happens, the real-time situation on the scene is often sent to the remote PC for storage. And the situation is notified to the user's PDA via SMS. In this way, even if the criminals escape by chance, they can be brought to justice through the photo information stored on the PC.

Light sensor module

Considering that the video acquisition effect of the camera is not good in a low-light environment, it is necessary to monitor the external environment through this module and adjust the light intensity within the camera range in time.

External energy-saving small lamp

Connect an energy-saving lamp via USB or serial port. When the light sensor detects that the light intensity is lower than a certain value, a control signal is generated to turn on the energy-saving lamp. Of course, you can also send a command to turn on or off the lamp to the system platform on a remote PC or PDA. It should be emphasized here that the remote control command has a high priority. Therefore, if the lamp is turned off at the remote end, even if the working environment becomes dark, it cannot be turned on by automatic detection.

3. System software architecture

This system uses C\C++ and assembly language as the system development programming language, and the program implements modular design. Finally, through the combination and coordination of various modules, a multifunctional embedded system is formed.

Development platform: windows xp\vista\7

Integrated Development Environment: AVR Studio 5

Programming language: C\C++ language, assembly language

Program download: Download via emulator

The software design should mainly include: uClinux system transplantation, hardware drivers of related interfaces, image preprocessing and filtering, image recognition, image compression and transmission, remote PC command reception and execution, etc.

uClinux transplantation

uClinux system is widely used in the field of microcontrollers, especially for CPUs without MMU (memory management unit), usually used for embedded operating systems with very small memory or Flash. Therefore, it meets the requirements of AVR32 AT32UC3A processors. (No MMU)

Implementation process: To connect the uClinux system, you first need to configure and program Uboot, generate the u-boot..hex file, and burn it into the flash through the development tool. Then configure and compile the uClinux kernel. After correctly configuring the kernel, recompile and download it to the target board.

Image preprocessing filtering

The original image acquired by the camera always has some noise. If the interference noise is too large, the image quality will be reduced or even drowned in the noise. Therefore, it is necessary to perform a pre-processing and filtering process on the image.

Image Identification

The image recognition here mainly includes the recognition of flames and faces. This is the focus of this project and requires the support of relevant algorithms.

Image compression transmission

Due to the limitation of storage space and reading transmission speed, the image needs to be compressed. Here we plan to use the JPEG compression algorithm, which is a lossy compression based on discrete cosine (DCT). The block diagram of the specific image compression encoding and decoding algorithm is as follows:

Remote PC (or PDA) command reception and execution

The command functions sent by the remote PC mainly include:

Turn on the camera

Turn off camera

Release stored data

Take 3 consecutive shots

Set the image resolution

Turn off energy-saving lights

4. System software process

5. Expected results of the system

First, install the camera connected to the system indoors and complete the relevant configuration.

During the day, when there are user members at home, the camera is turned off to save energy. When all family members are out, they send a command to the system through mobile devices such as mobile phones or computers to request to turn on the camera. Then the system starts monitoring. When it is rainy, the system can automatically adjust whether to turn on the energy-saving lamp according to the indoor light intensity.

When an abnormal situation occurs, such as a fire at home, the system will monitor whether there is anyone at the fire scene to confirm whether it is arson or someone is trapped indoors. If there is someone at the scene, the scene picture will be saved on the remote PC to facilitate the investigation of the cause of the fire afterwards.

Then, a text message will be sent to the user's mobile phone to inform them of the fire danger, and the camera will be turned off (you can choose to automatically dial 119 first). If there is no fire, when all family members are not at home, and a stranger breaks in, the camera will detect the person's appearance, take a picture of the evidence and send it to the remote PC, and send a text message to remind the user that the house has been burglarized, so the user can report the crime in time.

When user members go home, they can use mobile phones and other devices to turn off the camera to avoid misjudgment. At night, energy-saving lamps are generally needed to be turned on to better monitor the work.