Sensor‐based inter‐ and intra‐row weed control methods in sugar beet, sunflower, and maize

Abstract

Sensor‐guidance and camera‐based weed detection systems have improved weed control, resulting in higher weed control efficacy (WCE), less herbicide use, and less crop damage. In this study, four sensor‐guided inter‐ and intra‐row weeding systems were tested in sugar beet, maize, and sunflower. Those four systems included (1) camera‐guided inter‐row hoeing (HOE), (2) HOE + intra‐row finger weeding (HOE+F), (3) HOE + camera‐guided intra‐row hoeing (HOE + InRow) and (4) HOE + intra‐row band spraying (HOE + BS). A broadcast herbicide treatment and an untreated plot were included as control. Six experiments were conducted in 2022 and 2023 in Southwestern Germany. Inter‐row and intra‐row weed density and crop density were assessed before and after treatment. Yield was measured for each plot. The systems were evaluated based on WCE, crop losses (CL), yield, and herbicide savings. All sensor‐based weeding systems controlled at least on average 77% of the weeds. HOE+BS achieved 91% WCE and was similar to the broadcast herbicide application (92% WCE). HOE on average controlled 90% of the inter‐row weeds but only 69% of the in‐row weeds. HOE and HOE+F had 14% less intra‐row WCE than inter‐row WCE. HOE+InRow resulted in 77% inter‐row and intra‐row WCE. Sensor‐guided weed control did not cause significant CL or yield reduction compared to the broadcast herbicide application, except for HOE+InRow in maize 2023 because of wrong setting in the segmentation and weed/crop classification algorithm. This study underlines that sensor‐guided hoeing and the combination of band spraying with inter‐row hoeing provide effective and robust alternatives to conventional broadcast herbicide application in row crops.