Menu

GPS-based weed control in vegetable field [Student]

 

Scientific Article

Researchers at UC Davis and elsewhere have demonstrated the technical feasibility of using RTK-GPS technology to automatically create crop plant maps at planting (Sun et al., 2010; Perez-Ruiz, 2011). This paper describes the development and field evaluation of a completely automatic system for intra-row weed control in tomato fields. This system uses an automatically generated GPS crop plant map to automatically control the path of a set of mechanical intra-row weed knives in order to kill weeds growing between tomato plants in the seedline. The manual labor savings were evaluated in controlled field trials and the labor savings benefits associated with the use of this system are reported.[1]

 

Photograph of the RTK-GPS automated transplant mapping system being used to plant tomato transplants on the UC Davis campus farm.
Figure 1: RTK-GPS automated transplant mapping system being used to plant tomato transplants on the UC Davis campus farm.
In total, 2278 tomato transplants were automatically mapped using RTK–GPS during planting in the sixteen rows of the test plot for this study. The mean weed density in the test plot was 23.9 weeds/m2 and no significant difference in weed load between treatments was observed (p–value=0.44). RTK–GPS Fixed quality was obtained for all GPS antenna positions recorded during transplanting and mapping performance was comparable to that observed by Sun et al., (2010). 
Diagram showing the three weeding zones: A=inter-row (orange dashes with < symbols), B=intra-row (gray dashes with small square inside), and C=safety zones and the ideal path of the weed knives
Figure 2: Diagram showing the three weeding zones: A=inter-row (orange dashes with < symbols), B=intra-row (gray dashes with small square inside), and C=safety zones and the ideal path of the weed knives.

 

The automatic intra-row weeding machine successfully killed intra-row weeds between tomato plants without a significant (p-value=0.57) reduction in crop plants. Video analysis of knife operation showed satisfactory dynamic performance in following the desired cutting path. A significant (α=0.01) reduction in the amount of man-hours required for follow-up hand weeding was observed in rows automatically weeded by the machine. Figure 3 shows boxplots for the distributions of the man-hours required for hand weeding in the test. In the control treatment a mean of 206.7 h ha-1 was required to hand weed the intra-row and close-to-crop zones. Although the boxplot for the control treatment appears asymmetric, the Shapiro-Wilks test failed to reject the assumption of a Normal distribution for either treatment, (p-values &gt; 0.1). The additional hand weeding required in region B in Figure 2. For the RTK GPS automatic weeding treatment following the automatic intra-row weed knife operation was 99.7 h ha-1. This data shows that the automatic system developed can replace a significant amount of hand hoeing labor currently required to perform weed control in areas between crop plants in the row centerline.

Overall, a 52% reduction in the man-hours per hectare was obtained by using the automatic GPS-based intra-row weeding machine. Assuming hand weeding labor costs of US$10 per hour, this level of labor reduction potentially represents a significant savings in the cost of manual labor for hand hoeing. The cost savings in hand labor will need to offset the equipment cost, making the system more economically advantageous in organic production
systems where weed loads tend to be highest. Utilization of RTK GPS equipment for other tasks in the crop production system can help disperse the equipment cost across many cultural practices, reducing the equipment cost penalty in the weed control operation and may make it economically viable in conventional production systems.

 

Comparison of intra-row hand weeding labor required for RTK GPS automated weed control vs. conventional hand hoeing.
Figure 3: Comparison of intra-row hand weeding labor required for RTK GPS automated weed control vs. conventional hand hoeing.

 

References

[1]Slaughter, David & Pérez-Ruiz, Manuel & Fathallah, Fadi & Upadhyaya, Shrini & Gliever, Chris & Miller, Brandon (2019), GPS-Based Intra-Row Weed Control System: Performance and Labor Savings.

Last posts

Sidebar