Systemic Crop Signaling for Automatic Recognition of Transplanted Lettuce and Tomato under Different Levels of Sunlight for Early Season Weed Control

Abstract

Conventional cultivation works to control weeds between the rows, but it ignores the weeds in crop rows which are most competitive with crops. Many vegetable crops still require manual removal of intra-row weeds not otherwise controlled by herbicides or conventional cultivation. The increasing labor costs of weed control and the continued emergences of herbicide-resistant weeds are threatening grower ability to manage weeds and maintain profitability. Intra-row weeders are commercially available but work best in low weed populations. One strategy for rapid weed crop differentiation is to utilize a machine-detectable compound to mark a crop. This paper proposes a new systemic plant signaling technology that can create machine-readable crops to facilitate the automated removal of intra-row weeds in early growth stages. Rhodamine B (Rh–B) is an efficient systemic compound to label crop plants due to its membrane permeability and unique fluorescent properties. The project involves applying solutions of Rh–B at 60 ppm to the roots of lettuce and tomato plants prior to transplantation to evaluate Rh–B persistence in plants under different levels of sunlight. Lettuce and tomato seedlings with the systemic Rh–B should be reliably recognized during their early growth stages. An intelligent robot is expected to be developed to identify the locations of plants based on the systemic signal inside. Reduced light treatments should help to alleviate the photodegradation of Rh–B in plants. After being exposed to full sunlight for 27 days, the systemic Rh–B would be detectable in tomato branches and lettuce ribs, and these plants are tolerant to root treatments with this fluorescent compound. This paper describes the project background and plan as well as the anticipated contributions of the research to allow the machine vision system to reliably identify the crop plants, and thus showing technical feasibility for outdoor weed control.

Automatic Evaluation of Wheat Resistance to Fusarium Head Blight Using Dual Mask-RCNN Deep Learning Frameworks in Computer Vision

“…Computer vision-based phenotyping is a rapid, high-throughput, and non-destructive technique for capturing many types of traits [13][14][15][16][17]. Imaging techniques such as hyperspectral imaging (HSI) [18], and red-green-blue (RGB) imaging [19], have been widely used to study the complex traits associated with plant growth, biomass, yield, and responses to biotic stresses such as disease and abiotic stresses such as cold, drought and salinity [20][21][22].…”

A real-time smart sensing system for automatic localization and recognition of vegetable plants for weed control

“…This technology creates a machine-readable signal on plants that allows the marked crops to be readily identified by a computer vision system. Crop signaling technology has been successfully used to identify different target plants in weeds (Nguyen et al, 2017;Vuong et al, 2017;Raja et al, 2019a;Su, 2020). For example, Raja et al (2020b) developed a device containing two cameras to automatically detect lettuce and weeds based on crop signaling.…”