DEVELOPMENT OF MACHINE VISION AND LASER RADAR BASED AUTONOMOUS VEHICLE

Views:
 
     
 

Presentation Description

DEVELOPMENT OF MACHINE VISION AND LASER RADAR BASED AUTONOMOUS VEHICLE GUIDANCE SYSTEM

Comments

Presentation Transcript

PowerPoint Presentation:

SEMINAR TOPIC: Development of Machine Vision and Laser Radar Based Autonomous Vehicle Guidance System for Citrus G rove N avigation Speaker : Ghotekar Ravikant Sainath (M.Tech 1 st year) Roll No. : 13AG61R16 Welcome Author: Thomos F. Burks & V. Subramainan (Computer and Electronics in Agriculture, June 2006)

PowerPoint Presentation:

Content 2

INTRODUCTION:

INTRODUCTION 3

INTRODUCTION ......continued:

INTRODUCTION ......continued 4

PowerPoint Presentation:

Objectives 5

PowerPoint Presentation:

Material & Methods Frame Grabber To send error info from PC to microcontroller Camera 6

PowerPoint Presentation:

Material & Methods ….Continued Tractor with all additional attachment for autonomous guidance 7

PowerPoint Presentation:

Material & Methods ….Continued System overall working 8

PowerPoint Presentation:

Machine Vision A bility of a computer to "see” Includes one or more video cameras for obtaining images for the computer to interpret With computer vision, there is always a need of physical feature like colour difference for the vision system to be able to sense effectively Vision involves many complicated algorithms for image processing and recognition Camera mounted at the front Threshold image 9 Material & Methods ….Continued

PowerPoint Presentation:

Laser Radar (Ladar) Principle: Time-of-flight Measurement R emote sensing technology that measures distance by illuminating a target with a  laser and analysing the reflected light Distance = (Speed of Light x Time of Flight) / 2 used for ranging and obstacle avoidance 10 Material & Methods ….Continued Ladar Mounted on top of the tractor

PowerPoint Presentation:

An artificial testing path of hay bales was made Algorithms for processing the image and ladar information had developed for citrus orchard environment & hay bales environment Experiment were conducted on both testing path & Citrus orchard environment by both below guidance system A) Vehicle Guidance System by Machine Vision B) Autonomous Guidance System by Laser Radar System Material & Methods ….Continued EXPERIMENTAL PROCEDURE 11

PowerPoint Presentation:

Material & Methods ….Continued EXPERIMENTAL PROCEDURE ON ARTIFICIAL TESTING PATH Two types of paths: Straight path & Curved path The hay bale width was 45 cm, length of the straight path was 70 feet & an extension of 30 feet was given to form a curved path The path width was 3.5 m throughout the length. Experiments conducted for three different speeds i.e. 1.8 m/s, 3.1m/s, 4.4m/s A rotating blade was attached to drawbar, which marked a line on the ground as the vehicle moved (path center traveled by the tractor) Manually error was measured Above procedure repeated to calculate the path root mean square error, standard deviation, maximum error and average error 12

PowerPoint Presentation:

Material & Methods ….Continued VEHICLE GUIDANCE SYSTEM BY MACHINE VISION Color: discriminator for segmenting the path Camera calibration: To convert pixel distance to true distance To account for the varying weather conditions : images collected over a period of 6 days in 2 months from morning to evening at half an hour intervals Three types of conditions observed Cloudy days: trees are darker than the path Bright sunny days: trees are darker than the path but all pixel intensity values are elevated Early morning and evening: when the sunlight causes the trees on one side of the row to be brighter than the path and the trees on the other side to be darker than the path Based on this database of images, a segmentation algorithm was developed 13

PowerPoint Presentation:

Flowchart : Algorithm for path finding for adaptive RGB threshold value using machine vision 14

PowerPoint Presentation:

Fig : Path boundary Fig : Tree canopy segmentation Fig : Raw image Fig. Machine vision results for citrus grove alleyway 15

PowerPoint Presentation:

Material & Methods ….Continued VEHICLE GUIDANCE BY Laser Radar Guidance System 16

PowerPoint Presentation:

PID: Proportional integral derivative controller: attempts to minimise the error by adjusting the process control inputs Material & Methods ….Continued DESIGN OF PID CONTROL FOR STEERING CONTROL 17

PowerPoint Presentation:

Error Calculation Desired position = (Right side tree boundary + Left side tree boundary) / 2 Error = Desired position – current position Material & Methods ….Continued FORMULAE USED Line fitting: Least square method Pixel Distance to actual Distance Conversion 100 cm = 177 pixels Distance of the tractor centre from the hay bales Distance = Radial Distance at the hay bale * cosine (Angle at that point) 18

PowerPoint Presentation:

Results & Discussion 19

PowerPoint Presentation:

Results & Discussion ….Continued Performance of machine vision guidance in the straight path @ 1.8 m/s @ 3.1 m/s @ 4.4 m/s 20

PowerPoint Presentation:

Results & Discussion ….Continued Performance of laser radar guidance in the straight path @ 4.4 m/s @ 3.1 m/s @ 1.8 m/s 21

PowerPoint Presentation:

Results & Discussion ….Continued Performance of laser radar guidance in the curved path Performance of machine vision guidance in the curved path @ 3.1 m/s @ 3.1 m/s 22

PowerPoint Presentation:

Conclusions Machine vision and laser radar based guidance systems were developed to navigate a tractor through the alleyway of a citrus grove A PID controller was developed and tested to control the tractor using the information from the machine vision system and laser radar It was found that the ladar-based guidance was the better guidance sensor for straight and curved paths at speeds of up to 3.1 m/s Machine vision-based guidance showed acceptable performance at all speeds and conditions The average errors were below 3 cm in most cases. The maximum error was not more than 6 cm in any test run Experiments demonstrated the accuracy of the guidance system under test path conditions and successful guidance of the tractor in a citrus orchard alleyway Additional testing is needed to improve the performance in the citrus orchard 23

PowerPoint Presentation:

References Subramanian, V., Burks, T.F., Singh, S., 2004. Autonomous greenhouse sprayer vehicle using machine vision and ladar for steering control. Appl.Eng. Agric. 21 (5), 935–943. Bell, T., Bevly, D., Biddinger, E., Parkinson, B.W., Rekow, A., 1998. Automatic tractor row and contour control on sloped terrain using Carrier-Phase Differential GPS. In: Proceedings of the Fourth International Conference on Precision Agriculture. Misao, Y., 2001. An image processing based automatic steering power system. In: Proceedings of the ASAE Meeting, California, USA. http://www.deere.com/en_US/careers/midcareer_jobs/field_robotics.html www.wikipaedia.com Gordon, G.P., Holmes, R.G., 1988. Laser positioning system for off-road vehicles. 24

PowerPoint Presentation:

25

PowerPoint Presentation:

26

Camera:

Camera Camera and its mount Camera mounted on the tractor Specification: Sony FCBEX780S CCD camera with analog video output format in NTSC (National Television System Committee standard) Camera was mounted at an angle of 45 degree to the horizontal 27

Frame Grabber:

Frame Grabber It converts the analog NTSC video signal to a digital 640 x 480 RGB bitmap image 28

Laser Radar (Ladar):

Laser Radar (Ladar) Sick LMS-200 ladar sensor It is a 180 degree one-dimensional sweeping laser which can measure at 1.0/0.5/0.25 degree increments with maximum range of up to 80 m Mounted on top of the tractor cab just below the camera positioned at 45 degree to the horizontal Laser radar Laser mounted on top of the tractor 29

Computer:

Computer 4 GHz Pentium4 processor running Windows 2000 pro operating system S oftware (to develop algorithms): Microsoft Visual C++ Computer, monitor and keyboard mounted in the cabin Computer 30

Microcontroller:

Microcontroller 586 Engine controller board with a P50 expansion board from TERN Inc. It is a C++ programmable controller board based on a 32-bit system Function : For executing Real time-time control of the Servo valve & Encoder feedback loop. 31 Amplifier: To scale the control voltage from the microcontroller to the servo valve

Encoder:

Encoder Stegmann Heavy Duty HD20 encoder Function: Feeding back the wheel angle to the control system Encoder Calibration Tractor was positioned at a place in the lab Angular positions were marked on the ground From the centre position, the steering wheel was rotated to get different angles of front wheel The number of pulses to reach different angles was noted (*…wheel centre was calibrated by trial and error) 32

Servo Valve:

Servo Valve 33

GPS Receiver:

GPS Receiver A GPS receiver was used to measure the vehicle displacement while conducting tests to determine the dynamics of the vehicle John Deere Starfire SF2000R Differential GPS receiver was used GPS mounted at the top of the tractor 34

Power Supply:

Power Supply Inverter: It supply required voltage to PC, the monitor and the laser radar Cigarette lighter power source The supply for the microcontroller and the hydraulic valve is taken from it Provided in tractor cabin 35

PowerPoint Presentation:

Radial Distance measured by Laser Radar 36

PowerPoint Presentation:

(a) (b) EXPERIMENTAL PROCEDURE ON ARTIFICIAL TESTING PATH Guidance system test path Fig. Curved path Fig. Straight path 37

PowerPoint Presentation:

Fig. Device used to mark the tractor route on the ground Fig. Marks on the ground indicating the path traversed EXPERIMENTAL PROCEDURE ON ARTIFICIAL TESTING PATH Path traced by the rotating blade 38

PowerPoint Presentation:

T H R E S H O L D E D I M A G E Shadow image Non shadow image 39

authorStream Live Help