| |
Web URL(s): | https://scisoc.confex.com/scisoc/2023am/meetingapp.cgi/Paper/151207 Last checked: 12/08/2023 |
Publication Type:
| Report |
Content Type: | Abstract or Summary only |
Author(s): | Gurjar, Bholuram;
Torres, Ubaldo;
Straw, Chase M.;
Bagavathiannan, Muthu |
Author Affiliation: | Gurjar: Presenting Author and Texas A&M University, College Station, TX; Torres and Bagavathiannan: Texas A&M University, College Station, TX; Straw: Horticultural Science, Texas A&M University, College Station, TX |
Title: | Developing a machine-learning based weed detection and localization framework for precision herbicide applications in turf |
Section: | 377 Other records with the "377" Section
Turf pest management poster: Diseases, insects, weeds I (includes student competition) Other records with the "Turf pest management poster: Diseases, insects, weeds I (includes student competition)" Section
C05 turfgrass science Other records with the "C05 turfgrass science" Section
|
Meeting Info.: | St. Louis, Missouri: October 29-November 1, 2023 |
Source: | ASA, CSSA, SSSA International Annual Meeting. 2023, p. 151207. |
Publishing Information: | [Madison, Wisconsin]: [American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America] |
# of Pages: | 1 |
Abstract/Contents: | "Weed control is an important part of turfgrass management, and the chemical method is both the most effective and economically viable. However, the widespread use of blanket herbicide applications has harmful effects on the environment and reduces the aesthetic look of the turfgrass. To mitigate these issues, targeted herbicide applications can reduce herbicide usage and minimize environmental harm. While various researchers have explored deep learning for weed detection in turfgrass, there remains a need to utilize unmanned aerial system (UAS) imagery for weed detection and localization, specifically targeting individual annual bluegrass (Poa annua L.) plants. In this research, we have developed a machine learning-based framework for weed detection and localization for precision herbicide application in turfgrass. Data was collected at different growth stages of annual bluegrass in Deer Park, Texas, with the specific objective of identifying and localizing annual bluegrass in turfgrass. We used the You Only Look Once (YOLO) deep learning model for weed detection, and a geotransformation function was used to convert image coordinates into global coordinates for site-specific herbicide applications to manage individual weed plants in turfgrass using drones, robots, and GNSS-guided boom sprayers. The results showed that the YOLOv7-w6 model had a detection accuracy of 78% compared to the YOLOv5l (68%), and YOLOv8l (66%). Once the model detects weeds in a photo, geotransformation transforms the center pixel of the bounding box into the latitude and longitude of each weed plant and achieves sub-centimeter-level accuracy; however, its performance depends on the accuracy of the orthophoto georeferencing. The developed model can be utilized for site-specific herbicide applications to manage individual annual bluegrass plants in turfgrass using drones, robots, and GNSS-guided boom sprayers. Future improvements will include real-time weed detection and testing the model's robustness at different growth stages and with various turf species." |
Language: | English |
References: | 0 |
See Also: | Updated version appears in Weed Science Society of America - Southern Weed Science Society Joint Meeting, 2024, p.23, with variant title,"A machine learning framework for the detection and localization of annual bluegrass (Poa annua) in bermudagrass turf", R=336047. R=336047 |
Note: | This item is an abstract only! |
| ASA/CSSA/SSSA Citation (Crop Science-Like - may be incomplete): Gurjar, B., U. Torres, C. M. Straw, and M. Bagavathiannan. 2023. Developing a machine-learning based weed detection and localization framework for precision herbicide applications in turf. Agron. Abr. p. 151207. |
| Fastlink to access this record outside TGIF: https://tic.msu.edu/tgif/flink?recno=333566 |
| If there are problems with this record, send us feedback about record 333566. |
| Choices for finding the above item: |
| Web URL(s): https://scisoc.confex.com/scisoc/2023am/meetingapp.cgi/Paper/151207 Last checked: 12/08/2023 |
| Request through your local library's inter-library loan service (bring or send a copy of this TGIF record) |