Design and Implementation of Intelligent Life Detection Robot
Electronic Information Science and Technology 6666666666 Name of student: Instructor:
Summary
Natural disasters such as earthquakes, landslides, and mudslides have severely affected and restricted the development of human society and economy. The post-disaster environment is complex and dangerous, and secondary disasters may occur at any time. These problems will hinder the smooth development of post-disaster rescue work. Based on this, an intelligent life detection robot based on the principle of infrared thermal imaging is designed. The robot uses FLIR Lepton thermal imaging camera to collect infrared thermal images, and OpenMV4 as the processing core to extract image features. The camera is installed on a two-degree-of-freedom pan-tilt to easily realize multi-directional detection. In order to realize remote monitoring, the control terminal can control the movement of the robot through the 433M wireless module. At the same time, the robot transmits the infrared thermal image to the FPV screen of the monitoring terminal in real time through the 5.8G wireless image transmission. The robot's motion system consists of mecanum wheels and a shock-absorbing structure, making it more flexible and rapid in rescue work. Finally, use MATLAB software to verify and simulate the function of the life detection robot. Perform non-uniformity correction on the collected original infrared images, image feature extraction, and realize the separation of living body and background. At the same time, the MATLAB toolbox is used to simulate the robot's intelligent obstacle avoidance algorithm. The results of debugging and simulation experiments show that the life detection robot can basically complete the target tasks, laying a foundation for future experimental research and technological improvement.
Keywords: life detection robot; infrared thermal imaging; mecanum wheel; intelligent obstacle avoidance
Abstract
Earthquake, landslide, debris flow and other natural disasters have seriously affected and restricted the development of human society and economy. After the disaster, the environment is complex and the risk is high, and secondary disasters will occur at any time. All these problems will hinder the smooth development of the disaster rescue work. Based on this, an intelligent life detection robot based on the principle of infrared thermal imaging is designed. The robot uses FLIR lepton thermal imaging camera to collect infrared thermal images, and openmv4 as the processing core to extract image features. The camera is installed on a two degree of freedom platform, which can easily realize multi-directional detection. In order to realize remote monitoring, the control terminal can control the robot movement through 433M wireless module. Meanwhile, the robot transmits the infrared thermal image to the FPV screen of the monitoring terminal in real time through 5.8G wireless image transmission. The robot’s motion system is composed of mcnami wheel and damping structure, which makes it more flexible and rapid in the rescue work. Finally, the function of life detection robot is verified and simulated by MATLAB software. The nonuniformity correction and the feature extraction of the original infrared image are carried out to separate the living body and the background. At the same time, matlab toolbox is used to simulate the robot intelligent obstacle avoidance algorithm. The results of debugging and simulation experiments show that the life detection robot can basically complete the target task, which lays a foundation for future experimental research and technical improvement.
Key words: life detection robot; infrared thermal imaging; mcnamm wheel; intelligent obstacle avoidance
1. Background and significance
Earthquakes, mine accidents, etc., disaster sites often have high-temperature radiation, toxic gases, etc., and even the Tianjin explosion accident may cause secondary disasters. Rescuers cannot immediately know whether there are victims at the disaster site or cannot determine the exact location and surrounding environment of the victims, which may delay the best rescue time. It is now necessary to design an intelligent life detection robot that can go deep into the first scene after the disaster to find out the situation for rescue work. It is of great significance to avoid or reduce the casualties of rescue personnel and improve rescue efficiency.
Second, the status quo at home and abroad
Developed countries such as the United States and Japan are the first to carry out research on life detection robots. Many universities and research institutions have developed life detection and rescue robots for different purposes. Our country started late in this field, and its research situation lags behind the world's advanced level. However, the detection robots studied at home and abroad are far from the requirements of practical applications. There are short communication distances, obstacle avoidance and obstacle crossing performance are average, and the environment with dense smoke or high dust concentration cannot determine the accuracy of survivors. Technical issues such as location.
3. Design content
The block diagram of the intelligent life detection robot design is shown in Figure 1. The whole system is divided into three parts, which are the infrared thermal imaging part, the robot main control part, and the robot remote control part. In the thermal imaging part, the control core is OpenMV4 H7 for feature extraction of the collected thermal images, and the FLIR Lepton 3.5 infrared camera is selected to collect infrared images. In order to realize remote monitoring, the thermal images are transmitted to the remote FPV monitor in real time through the 5.8G wireless image transmission module. The main control part of the robot uses STM32F405 as the processing core to collect and process sensor information, control the motion system of the robot, and realize data communication with the robot remote control through the 433M wireless module.
Figure 1 Block diagram of the design of intelligent life detection robot
The main content of this design includes three parts: the hardware design of the intelligent life detection robot, the program design and the simulation of the infrared image processing algorithm by MALTAB.
The hardware design part includes the thermal imaging camera circuit, the robot main control circuit, and the remote control circuit. It includes the design of the overall hardware scheme, the selection of the main control chip, the selection of the involved smart sensors, the system power circuit, the motor drive circuit, and the Design of display circuit for debugging. Program design part, including the main control part of the robot program design: motor driver, sensor data acquisition program and processing program, wireless communication program, etc.; infrared thermal imaging part program design: infrared thermal image acquisition program, thermal image feature extraction program, 5.8G real-time wireless image transmission program, etc.; remote control part program design: wireless digital transmission program, push rod and button control program, gyroscope sensor data acquisition and processing program. The MATLAB simulation part includes feature extraction of collected infrared images, simulation and improvement of infrared image processing algorithms, and simulation of robot intelligent obstacle avoidance algorithms.
The processing of infrared images is the core of the design. Due to the non-uniformity of the uncooled infrared detection camera and the special working environment of the robot, it is necessary to perform non-uniformity correction and denoising processing on the collected infrared images, and then perform feature extraction on the image , Separate life form and background. The improved BP neural network algorithm is used to correct the non-uniformity of the original infrared image, the noise is simulated by MATLAB software, and the denoising algorithm is designed. In order to make the detection target more obvious, the denoised image is image enhanced, and the Sobel operator is finally used Perform edge detection to achieve separation of target and background.
Four, design results
The hardware design part has completed the production of three circuit boards, namely the robot main control board, the motor drive board, and the robot remote control board. The software part completes the robot motion system control program, remote control program, infrared image acquisition program and thermal image feature extraction program. The MATLAB simulation part completes the non-uniformity correction of the original infrared image, image filtering and image enhancement processing, and the edge extraction of the thermal image.
As shown in Figure 2, the original image is grayed out, and then salt and pepper noise and Gaussian noise are added. Median filtering and mean filtering are used to filter the noised image, and the histogram equalization algorithm is used for the denoised image. Image enhancement processing.
Figure 2 Infrared image processing results
After image processing, the edges and gray jumps of the image are enhanced to make the image clear and make sufficient preparations for the subsequent image edge extraction. As shown in Fig. 3, six kinds of operator processing image algorithms are written with MATLAB software to detect the edges of the enhanced image respectively.
Figure 3 Comparison of edge detection results
By comparing the processing effects of the six edge detection operators, it can be seen that the log operator and the Canny operator are more suitable for the edge detection processing of this infrared image, and the mathematical morphology method is more effective, and the edge of the background image is also outlined. Therefore, when applied to robots, mathematical morphology may be more suitable. According to different situations, other algorithms are also needed to process infrared images, which needs to be improved.
Five, summary and experience
The results of debugging and simulation experiments show that the life detection robot can basically complete the target task. After the collected infrared thermal images are added with simulated noise in the laboratory, the image features are extracted. The boundary between the living body and the background is obvious, and the living body detection in the post-disaster environment can be realized. Although there are still some shortcomings in the intelligent life detection robot, it has basically achieved the expected design goal, laying a foundation for future experimental research and technological improvement. In the specific design process, I also encountered many problems, and after continuous efforts, the difficulties were overcome little by little. Through this graduation project, I not only became more familiar with the design steps and methods of engineering systems, and mastered a lot of scientific and technological knowledge in image processing and robot control from the perspective of application, but also clarified how to think, optimize time, and do things efficiently. The basic principle.
This graduation project has given me a lot of new knowledge and honed my will. It is not only a learning process of engineering design, but also a test of my previous learning achievements. It also provides some valuable for my future study and work. experience.
references
[1] Cheng Pei. Intelligent detection robot system design [D]. Harbin Engineering University, 2013.
[2] Ning Weizhe. Design of mobile detection robot based on lidar [D]. Shandong University, 2018.
[3] Alex Ellery. Environment robot interaction the basis for mobility in Planetary micro-rovers(J). Robotics and Autonomous Systems, 2004, 7(5): 1-10
[4] Tao
Sun. Life detection system based on infrared thermal imaging [D]. Yanshan University, 2011. [5] Teleoperation of a Mobile Robot Using a Force-Reflection Joystick With Sensing Mechanism of Rotating Magnetic Field, 2010, 15(1)
[6] Wang Feng. Research on the application of communication and audio and video systems for underground exploration robots (D). Master's degree thesis of Shandong University. 2009.5:2-3.
[7] Wei Jurong. Design of simulation platform for motion performance of environment detection robot (D). Master's degree thesis of Xidian University. 2010.1:3-4.
[8] Cao Yanpeng, Xu Baobei, He Zewei, etc. Infrared thermal imaging signal processing technology Research progress[J]. Vibration, Testing and Diagnosis, 2018, 038(002).
[9] Vollmer M,Mollmann K P.Infrared thermal imaging (J).European Journal of Physics,2010,32(5):8- 25.
[10] Xu Zhiying, Li Jinping. MATLAB and its application in image processing [J]. Computer and Modernization, 2003 (4): 64-65.
[11] Lan Jie, Zhang Haoran. Based on BP neural network QR code area Extraction[J]. Microcomputers and Applications, 2015(1):50-52.
Reply PPT
