Document Type : Research Article
Authors
1 Mechanics of Biosystems Engineering, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran
2 Tarbiat Modares University
3 Agricultural Sciences and Natural Resources University of Khuzestan
4 Department of Agriculture, Khuzestan, Agricultural Science and Natural Resources University, Khuzestan Iran
Abstract
Introduction
Today, machine vision systems are increasingly used in agriculture. The use of this technology in this field can help preserve agricultural resources while reducing manual labor and production costs. In the field of agricultural automation, the accurate detection of crop rows is recognized as an important and challenging issue in terms of weed identification and automatic guidance of machines, and it is necessary to examine practical solutions in order to optimize it. Therefore, the purpose of this study is to accurately identify the basil cultivation rows in order to automatically route a robot in the cultivation field.
Materials and Methods
In one stage of this research, by taking six images in each growth period (third week, fourth week, and fifth week), weeds were removed between the crop rows; For this purpose, three different methods (area opening, dimensional removal and masking) were used. In another step, six images of crop rows without weeds were examined. Then, by performing image processing operations and implementing several routing algorithms (algorithms based on Hough transform, wavelet transform, Gabor filter, linear regression and the proposed algorithm of this study) on the images, the output of each of these algorithms compared to the specified ideal path by user was investigated. For this purpose, after capturing the image, all green areas were extracted from the image by performing the segmentation process. The weeds between crop rows were removed using three different methods during the growth period. In the next step, by applying each of the routing algorithms on the image, plant cultivation lines were identified and their equations were determined. Finally, the performance of the designed robot was evaluated using the most appropriate routing algorithm.
Results and Discussion
After examining the performance of three different methods of weed removal in three periods of plant growth (third week, fourth week and fifth week), it was shown that in all periods of plant growth, the masking method had output with the lowest error rate compared to the ideal path, the shortest operation time (1.64 seconds on average) followed by the dimensional removal and the area opening methods. In the following work, by carefully comparing the routes detected by different routing algorithms compared to the ideal routes and according to the results of t-test at 5% probability level, the superiority of the studied routing methods was determined as follows: Proposed method, Gabor filter method, linear regression method, Hough transform method and wavelet transform method. Finally, the proposed algorithm with the highest rate of adaptation to the ideal path (with an average error of 3.65 pixels) and the shortest operation time (4.79 seconds) was selected as the most appropriate routing algorithm, and the performance of designed robot was evaluated using it.
Conclusion
After measuring and comparing the execution speed and error rate of each of the studied routing algorithms, according to the results, it was determined that the proposed method, Gabor filter method, linear regression method, Hough transform method and wavelet transform method are preferable to each other in the field of routing, respectively. Finally, it was shown that the designed robot using the proposed algorithm (with an average error of 3.65 pixels) has desired performance.
Acknowledgment
The authors express appreciation for the financial support provided by Tarbiat Modares University.
Keywords
Main Subjects
)
Send comment about this article