Plant Counting of Cotton from UAS Imagery Using Deep Learning-Based Object Detection Framework

Sungchan Oh,Nothabo Dube,Jinha Jung,Murilo Maeda,Juan Landivar,Daniel Gonzalez,Anjin Chang,Akash Ashapure

doi:10.3390/rs12182981

Sungchan Oh, Nothabo Dube + Show 6 more

Open Access

https://doi.org/10.3390/rs12182981

Copy DOI

Abstract

Assessing plant population of cotton is important to make replanting decisions in low plant density areas, prone to yielding penalties. Since the measurement of plant population in the field is labor intensive and subject to error, in this study, a new approach of image-based plant counting is proposed, using unmanned aircraft systems (UAS; DJI Mavic 2 Pro, Shenzhen, China) data. The previously developed image-based techniques required a priori information of geometry or statistical characteristics of plant canopy features, while also limiting the versatility of the methods in variable field conditions. In this regard, a deep learning-based plant counting algorithm was proposed to reduce the number of input variables, and to remove requirements for acquiring geometric or statistical information. The object detection model named You Only Look Once version 3 (YOLOv3) and photogrammetry were utilized to separate, locate, and count cotton plants in the seedling stage. The proposed algorithm was tested with four different UAS datasets, containing variability in plant size, overall illumination, and background brightness. Root mean square error (RMSE) and R2 values of the optimal plant count results ranged from 0.50 to 0.60 plants per linear meter of row (number of plants within 1 m distance along the planting row direction) and 0.96 to 0.97, respectively. The object detection algorithm, trained with variable plant size, ground wetness, and lighting conditions generally resulted in a lower detection error, unless an observable difference of developmental stages of cotton existed. The proposed plant counting algorithm performed well with 0–14 plants per linear meter of row, when cotton plants are generally separable in the seedling stage. This study is expected to provide an automated methodology for in situ evaluation of plant emergence using UAS data.

Highlights

In cotton (Gossypium hirsutum L.) production, early crop establishment is a critical stage that determines the yield potential of a given field
From 300 labeled images of each flight, training and testing images were randomly chosen in a 2-to-110sopfl2it3, and mean average precision (mAP) was investigated as a measure of detection accuracy
While this study focuses on optimizing the three major hyperparameters to minimize the Root mean square error (RMSE) of plant cRoemuontte,Seuns.e20o2f0, a12,n2o98v1el object detection algorithm that is capable of detecting small objects18aonf 2d2 separating overlapping objects will greatly improve the performance of image-based plant counting

Summary

Introduction

In cotton (Gossypium hirsutum L.) production, early crop establishment is a critical stage that determines the yield potential of a given field. Yield potential is initially limited by the density of cotton plants and may be further impacted by unfavorable weather conditions, as well as disease and insect pressure [1,2,3,4]. One of the most common ways to evaluate plant density is manually counting the number of plants in 1/1000th of an acre (4047 m2) in the field [7]. The procedure should be repeated approximately 10 times in different locations and patterns to reduce sampling bias. This method is simple and straightforward, sampling bias can exist when the collected samples are not representative with respect to sampling location, and within-field variability is too high. Manual plant population assessment is time-consuming and labor-intensive

Methods

Results

Conclusion