Robotic automation to minimize the harsh working conditions and reduce the risk of work can satisfy the objectives of agricultural automation with subtle controlling and iection. About 10% of greenhouses in Iran produce cut flower rose, more than 600 producers. However, using robots have several limitations such as the possibility of collision robot with flowers and greenhouse structures, robot speed, robot ability to cargo traortation, and its compatibility with temperature and humidity conditions. The aim of this study was to design and build a robotic system to navigate between flower rows and detected ripe roses in the Isfahan University of Technology. Three-dimensional modeling of parts was done by CAE engineering and SolidWorks. To crossing specific path, a line follower robot was designed. Two separate algorithms for red and white-pink lip roses were written in MATLAB software to detect roses in real time. Communication between line following robot and rose detection algorithm in MATLAB software was performed with serial port. As long as the line following robot motors are turned on, the process of capture and image processing will not be done. When the motors were switched off, the detection operation was begun after one-second pause. At the end of detection operation, a command was sent to line following robot processor to start motors automatically. Initial evaluations color spaces available showed the OHTA color space with least disruption can separate red rose from complicated background in the greenhouse. This color space was not appropriate for the detection of white and pink lip roses. So, the method of red monochrome layer instead of compound layers was used to detect of white- pink lip roses. Success rate of red rose detection was 91% and 90% for white-pink lip. In order to estimate the number of roses in the greenhouse, two cameras were used in the horizontal and vertical alignment. Using the horizontally camera wasn’t appropriate because of observations of other rows of flowers. The rotation mechanism of vision system with 85% success rate placed camera in the position of the next row at the end of every row. Line following robot performance on the straight-line, moving around and stop at the stations was successful. Excessive height of roses, lack of exposure in the camera's view, light reflection from the leaves of white-pink lip roses, accumulation of red leaves from red rose, hiding ripe flowers under the leaves, and the flowers at the end or start decreased the success rate of its detection. Keywords: Robotics, Greenhouse, Cut flower, Image processing, Navigation.
