Pesticide-Free Robotic Control of Aphids as Crop Pests

Abstract

Because our civilization has relied on pesticides to fight weeds, insects, and diseases since antiquity, the use of these chemicals has become natural and exclusive. Unfortunately, the use of pesticides has progressively had alarming effects on water quality, biodiversity, and human health. This paper proposes to improve farming practices by replacing pesticides with a laser-based robotic approach. This study focused on the neutralization of aphids, as they are among the most harmful pests for crops and complex to control. With the help of deep learning, we developed a mobile robot that spans crop rows, locates aphids, and neutralizes them with laser beams. We have built a prototype with the sole purpose of validating the localization-neutralization loop on a single seedling row. The experiments performed in our laboratory demonstrate the feasibility of detecting different lines of aphids (50% detected at 3 cm/s) and of neutralizing them (90% mortality) without impacting the growth of their host plants. The results are encouraging since aphids are one of the most challenging crop pests to eradicate. However, enhancements in detection and mainly in targeting are necessary to be useful in a real farming context. Moreover, robustness regarding field conditions should be evaluated.

Agricultural Robot-Centered Recognition of Early-Developmental Pest Stage Based on Deep Learning: A Case Study on Fall Armyworm (Spodoptera frugiperda)

“…Linear and quadratic discriminant analysis and support vector machines were also experimented on, but they are required to attain precision. Lucet et al [ 21 ] proposed a mobile robot for pesticide-free aphid crop pest control with a ZED mini RGB-D sensor; this sensor provides stereo camera images for the YOLOv4 trained on a unique aphid dataset. Upon detection on the 2D RGB image (XY), the ZED mini SDK is used to localize the aphid pest approximately in 3D space (XYZ), allowing visual serving control to guide the proposed laser-based neutralization precisely.…”

“…Detecting insects is quite challenging, considering their tiny sizes for detection algorithms and their agility at the adult stage for precise neutralization. Lucet et al [ 21 ] also experimented with stereo sensor mount distances from ground and output image size sub-sampling to arrive at <30 cm and 800 × 800 pixels in 4 or 16 sub-parts for 0.73 sensitivity, respectively. Consequently, some researchers have employed an indirect method for controlling insects by limiting the growth of weeds on farms.…”

“…As a result, we suggest employing a deep-learning-based method to strategically hunt the insect pests’ larvae stage to serve as a means for simple robot configuration. Similar to [ 21 ], we also suggested using an off-the-shelf stereo sensor for deep-learning insect pest recognition and localization. The strategy takes the merit of the robustness of two deep-learning algorithms, a deep-learning classifier and a deep-learning detection model, to classify and detect insect pests at the early stage (i.e., larvae) during preplanting, as well as to precisely provide localization for the agrobot to localize above the insect pest to neutralize with any means that minimizes the quantity of chemicals to be used.…”

An optical system to detect, surveil, and kill flying insect vectors of human and crop pathogens

“…In lab tests, laser pulses were shown to be capable of killing flying mosquitoes ( Anopheles stephensi , Culex pipiens ), Asian citrus psyllid ( Diaphorina citri ) and, spotted wing drospophila ( Drosophila suzukii ) 127 – 131 . Aphids, which are among the most important vectors of plant viruses, are vulnerable to laser pulses while they feed on plants 134 , 135 . In this system, a robot bearing the photonic system moves down a row of crop plants, interrogates each plant, and pulses a laser to individual aphids attached to the stems and leaves.…”

Aphid Recognition and Counting Based on an Improved YOLOv5 Algorithm in a Climate Chamber Environment

“…To our knowledge, there have been few limited reports of aphid recognition using the YOLO algorithm. For example, Lacotte et al evaluated the performance of YOLOv4 using precision and recall criteria, and applied YOLOv4 for the detection of aphids in the pesticide-free robotic control process [34]. Xu et al proposed a lightweight SSV2-YOLO model based on YOLOv5s for the detection of sugarcane aphids in unstructured natural environments and the method reduced the model size [35].…”