Skip to main content
ARS Home » Plains Area » Temple, Texas » Grassland Soil and Water Research Laboratory » Research » Publications at this Location » Publication #420196
Research Project: Development of Enhanced Tools and Management Strategies to Support Sustainable Agricultural Systems and Water Quality

Location: Grassland Soil and Water Research Laboratory

Title: An innovative UAV-based approach for estimating crop stand counts amidst weed infestation

Author
item BAATH, GURJINDER - Texas Agrilife Research
item BAWA, ARUN - Texas Agrilife Research
item SAPKOTA, BALA RAM - Texas Agrilife Research
item Flynn, Kyle
item SAKAR, SAYANTAN - Texas Agrilife Research
item Smith, Douglas

Submitted to: Smart Agricultural Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/19/2025
Publication Date: 5/20/2025
Citation: Baath, G.S., Bawa, A., Sapkota, B., Flynn, K.C., Sakar, S., Smith, D.R. 2025. An innovative UAV-based approach for estimating crop stand counts amidst weed infestation. Smart Agricultural Technology. https://doi.org/10.1016/j.atech.2025.101030.
DOI: https://doi.org/10.1016/j.atech.2025.101030

Interpretive Summary: Unmanned aerial vehicles (UAV)-based imagery offers a rapid, cost-effective solution for estimating stand counts in row crops. However, most existing algorithms assume clean crop rows and do not account for common weed infestation. This research developed an image processing pipeline to estimate crop stand counts in both high and low weed pressure conditions. Results showed that the methodology was highly effective for stand count estimations in both high and low weed pressure situations. This novel standardized and scalable approach holds significant potential for improving UAV-based stand count assessments across various crops, particularly in organic and large-scale production systems.

Technical Abstract: Unmanned aerial vehicles (UAV)-based imagery offers a rapid, cost-effective solution for estimating stand counts in row crops. However, most existing algorithms assume clean crop rows and do not account for common weed infestation. This research developed an image processing pipeline to estimate crop stand counts in both high and low weed pressure conditions. The methodology was initially developed and tested on cotton using UAV imagery collected at three ground sampling distances (GSD) of 1.4, 2.1, and 2.8 cm, across three post-emergence stages: 9, 13, and 21 days after emergence (DAE) in 2022. The approach was later validated on maize during the 2023-2024 growing seasons. The pipeline integrated edge detection, row detection, and geometrical properties while minimizing reliance on spectral characteristics. Results showed that the methodology was highly effective for stand count estimations in both high and low weed pressure situations. Highest accuracies were achieved at low GSDs and early DAE stages, with an R2 of 0.80-0.85 between 9-13 DAE at 1.4 cm GSD for heavily weed infested cotton, and R2 of 0.94-0.97 between 12-17 DAE at 1.8 cm GSD for maize under minimal weed infestation. Flights at higher GSD (2.1 cm) performed reasonably well (R2 = 0.76) at early stages but showed a decline in accuracy with aggressive weed growth and overlapping cotton seedlings beyond 9 DAE. This novel standardized and scalable approach holds significant potential for improving UAV-based stand count assessments across various crops, particularly in organic and large-scale production systems.