Weed Images Research Articles

Dataset of annotated food crops and weed images for robotic computer vision control.

Weed management technologies that can identify weeds and distinguish them from crops are in need of artificial intelligence solutions based on a computer vision approach, to enable the development of precisely targeted and autonomous robotic weed management systems. A prerequisite of such systems is to create robust and reliable object detection that can unambiguously distinguish weed from food crops. One of the essential steps towards precision agriculture is using annotated images to train convolutional neural networks to distinguish weed from food crops, which can be later followed using mechanical weed removal or selected spraying of herbicides. In this data paper, we propose an open-access dataset with manually annotated images for weed detection. The dataset is composed of 1118 images in which 6 food crops and 8 weed species are identified, altogether 7853 annotations were made in total. Three RGB digital cameras were used for image capturing: Intel RealSense D435, Canon EOS 800D, and Sony W800. The images were taken on food crops and weeds grown in controlled environment and field conditions at different growth stages

Automated Weed Detection Systems: A Review

A weed plant can be described as a plant that is unwanted at a specific location at a given time. Farmers have fought against the weed populations for as long as land has been used for food production. In conventional agriculture this weed control contributes a considerable amount to the overall cost of the produce. Automatic weed detection is one of the viable solutions for efficient reduction or exclusion of chemicals in crop production. Research studies have been focusing and combining modern approaches and proposed techniques which automatically analyze and evaluate segmented weed images. This study discusses and compares the weed control methods and gives special attention in describing the current research in automating the weed detection and control.
 Keywords: Detection, Weed, Agriculture 4.0, Computational vision, Robotics

Weed Detection for Selective Spraying: a Review

Weed detection systems are important solutions to one of the existing agricultural problems—unmechanized weed control. Weed detection also helps provide a means of reducing or eliminating herbicide use, mitigating agricultural environmental and health impact, and improving sustainability. Deep learning-based techniques are replacing traditional machine learning techniques to detect weeds in real time with the development of new models and increasing computational power. More hybrid machine learning models are emerging, utilizing benefits from different techniques. More large-scale crop and weed image datasets are available online now, and this provides more data and opportunities for researchers and engineers to join and contribute to this field. This article provides a mini-review of all the different emerging and popular weed detection techniques for selective spraying, and summarizes the trends in this area in the past several years.

A Novel Crop Weed Recognition Method Based on Transfer Learning from VGG16 Implemented by Keras

This study used a method based on convolutional neural network model, VGG16, to identify images of weeds in the field. As the basic network, VGG16 has very good classification performance, and the network structure is unconventional. It is relatively easy to modify. It can fine-tune other data sets on this basis. Therefore, the transfer learning method is applied to our own Kaggle competition website. Download the weed data set. The site covers approximately 3, 500 images in 12 categories. Due to limited data and computational power, our model fixes the first 14 layers of VGG16 parameters for layer-by-layer automatic extraction of features, adding an average pooling layer, convolution layer, Dropout layer, fully connected layer, and softmax for classifiers. The layer has a total of 5 layers, for a total of 19 layers. The experimental results show that the final model performs well in the classification effect of 12 weed images. The accuracy rate on the training set is 98.99%, and the accuracy on the verification set is 91.08%. It can be applied to crop weed identification. It provides accurate and reliable judgment basis for positioning and quantitative chemical pesticide spraying, and is the key to achieving refined agriculture.

Combing K-means Clustering and Local Weighted Maximum Discriminant Projections for Weed Species Recognition

Abstract: Weed species identification is the premise to control weeds in smart agriculture. It is a challenging topic to control weeds in field, because the weeds in field are quite various and irregular with complex background. An identification method of weed species in crop field is proposed based on Grabcut and local discriminant projections (LWMDP) algorithm. First, Grabcut is used to remove the most background and K-means clustering (KMC) is utilized to segment weeds from the whole image. Then, LWMDP is employed to extract the low-dimensional discriminant features. Finally, the support vector machine (SVM) classifier is adopted to identify weed species. The characteristics of the method are that (1) Grabcut and KMC utilize the texture (color) information and boundary (contrast) information in the image to remove the most of background and obtain the clean weed image, which can reduce the burden of the subsequent feature extraction; (2) LWMDP aims to seek a transformation by the training samples, such that in the low-dimensional feature subspace, the different-class data points are mapped as far as possible while the within-class data points are projected as close as possible, and the matrix inverse computation is ignored in the generalized eigenvalue problem, thus the small sample size (SSS) problem is avoided naturally. The experimental results on the dataset of the weed species images show that the proposed method is effective for weed identification species, and can preliminarily meet the requirements of multi-row spraying of crop based on machine vision.

Maize seedling detection under different growth stages and complex field environments based on an improved Faster R–CNN

This paper presents an improved Faster R–CNN model for a field robot platform (FRP) aimed at automatically extracting image features and quickly and accurately detecting maize seedlings during different growth stages under complex field operation environments, with the goal of preparing for intelligent inter-tillage in maize fields. A FRP with five industrial USB cameras for data collection was used to capture a large number of sample images. The shooting angle range of the industrial USB cameras is 0–90°. The photographs were used to create an image database containing twenty thousand images of soil, maize and weeds. Ten selected pretrained networks were used to replace the network of the CNN feature computing component of the classic Faster R–CNN. A Faster R–CNN with VGG19 processed by the pretrained networks method is proposed. The Faster R–CNN algorithm used in this work represents a deep learning architecture that distinguishes maize seedlings and weeds under three field conditions: Full-cycle, Multi-weather and Multi-angle. This work achieved greater than 97.71% precision in the detection of maize seedlings with respect to soil and weeds. The precision rate of six-leaf to seven-leaf maize seedlings was 2.74% lower than that of the total test set. The precision rate under sunny conditions was 1.97% lower than that of the total test set. The precision rate of an angle shot of 0° was 0.95% lower than that of the total test set. The proposed model has significant potential for autonomous weed and maize classification under actual operating conditions.

The coreset variational Bayes (CVB) algorithm for mixture analysis

The pressing need for improved methods for analysing and coping with big data has opened up a new area of research for statisticians. Image analysis is an area where there is typically a very large number of data points to be processed per image, and often multiple images are captured over time. These issues make it challenging to design methodology that is reliable and yet still efficient enough to be of practical use. One promising emerging approach for this problem is to reduce the amount of data that actually has to be processed by extracting what we call coresets from the full dataset; analysis is then based on the coreset rather than the whole dataset. Coresets are representative subsamples of data that are carefully selected via an adaptive sampling approach. We propose a new approach called coreset variational Bayes (CVB) for mixture modelling; this is an algorithm which can perform a variational Bayes analysis of a dataset based on just an extracted coreset of the data. We apply our algorithm to weed image analysis.

Fully convolutional network for rice seedling and weed image segmentation at the seedling stage in paddy fields.

To reduce the cost of production and the pollution of the environment that is due to the overapplication of herbicide in paddy fields, the location information of rice seedlings and weeds must be detected in site-specific weed management (SSWM). With the development of deep learning, a semantic segmentation method with the SegNet that is based on fully convolutional network (FCN) was proposed. In this paper, RGB color images of seedling rice were captured in paddy field, and ground truth (GT) images were obtained by manually labeled the pixels in the RGB images with three separate categories, namely, rice seedlings, background, and weeds. The class weight coefficients were calculated to solve the problem of the unbalance of the number of the classification category. GT images and RGB images were used for data training and data testing. Eighty percent of the samples were randomly selected as the training dataset and 20% of samples were used as the test dataset. The proposed method was compared with a classical semantic segmentation model, namely, FCN, and U-Net models. The average accuracy rate of the SegNet method was 92.7%, whereas the average accuracy rates of the FCN and U-Net methods were 89.5% and 70.8%, respectively. The proposed SegNet method realized higher classification accuracy and could effectively classify the pixels of rice seedlings, background, and weeds in the paddy field images and acquire the positions of their regions.

Low-Power and High-Speed Deep FPGA Inference Engines for Weed Classification at the Edge

Deep neural networks (DNNs) have recently achieved remarkable performance in a myriad of applications, ranging from image recognition to language processing. Training such networks on graphics processing units (GPUs) currently offers unmatched levels of performance; however, GPUs are subject to large-power requirements. With recent advancements in high-level synthesis (HLS) techniques, new methods for accelerating deep networks using field programmable gate arrays (FPGAs) are emerging. FPGA-based DNNs present substantial advantages in energy efficiency over conventional CPU- and GPU-accelerated networks. Using the Intel FPGA software development kit (SDK) for OpenCL development environment, networks described using the high-level OpenCL framework can be accelerated targeting heterogeneous platforms including CPUs, GPUs, and FPGAs. These networks, if properly customized on GPUs and FPGAs, can be ideal candidates for learning and inference in resource-constrained portable devices such as robots and the Internet of Things (IoT) edge devices, where power is limited and performance is critical. Here, we introduce GPU- and FPGA-accelerated deterministically binarized DNNs, tailored toward weed species classification for robotic weed control. Our developed networks are trained and benchmarked using a publicly available weed species dataset, named DeepWeeds, which include close to 18 000 weed images. We demonstrate that our FPGA-accelerated binarized networks significantly outperform their GPU-accelerated counterparts, achieving a>7-fold decrease in power consumption, while performing inference on weed images 2.86 times faster compared to our best performing baseline full-precision GPU implementation. These significant benefits are gained whilst losing only 1.17% of validation accuracy. In this paper, this is a significant step toward enabling deep inference and learning on IoT edge devices, and smart portable machines such as agricultural robots, which is the target application.

Weed Growth Stage Estimator Using Deep Convolutional Neural Networks.

This study outlines a new method of automatically estimating weed species and growth stages (from cotyledon until eight leaves are visible) of in situ images covering 18 weed species or families. Images of weeds growing within a variety of crops were gathered across variable environmental conditions with regards to soil types, resolution and light settings. Then, 9649 of these images were used for training the computer, which automatically divided the weeds into nine growth classes. The performance of this proposed convolutional neural network approach was evaluated on a further set of 2516 images, which also varied in term of crop, soil type, image resolution and light conditions. The overall performance of this approach achieved a maximum accuracy of 78% for identifying Polygonum spp. and a minimum accuracy of 46% for blackgrass. In addition, it achieved an average 70% accuracy rate in estimating the number of leaves and 96% accuracy when accepting a deviation of two leaves. These results show that this new method of using deep convolutional neural networks has a relatively high ability to estimate early growth stages across a wide variety of weed species.

Development of vegetable intelligent farming device based on mobile APP

Poor realization of agronomic standards, low level of automation, time-consuming and labor-intensive farming are the main problems in traditional production process of vegetable. In order to improve vegetable cultivation intelligent and intensive level,reduce waste of production resources, an intelligent vegetable cultivation device was designed based on APP control, internet communications and image recognition technology, with the functions of remote control, precision sowing, quantitative dosing of liquid materials and weed recognition. The device mainly includes farming executing part, image processing part, STM32 microcontroller, and the APP which sends a command and control the device to work on the corresponding farming work. The precision seeding sowing worked through the gantry positioning and the cooperation of magnetic coupling tools and gas-suction sowing tools. By controlling the pump running time and monitoring flow volume with PVDF pressure sensor installed in the liquid pipeline interface, the liquid material was delivered quantitatively. Weed images were collected by CCD camera and recognition algorithm was developed based on BP neural network to obtain the weed location information. The test results show that the error rate of sowing was 2.75%, the passing rate of the plant spacing was up to 97.2%, no miss sowing occurs during the test; the error of the liquid delivery was within ± 5.8 g; true positive rate, true negative rate and accuracy of weed recognition were above 95%.

Fusion of pixel and object-based features for weed mapping using unmanned aerial vehicle imagery

The developments in the use of unmanned aerial vehicles (UAVs) and advanced imaging sensors provide new opportunities for ultra-high resolution (e.g., less than a 10 cm ground sampling distance (GSD)) crop field monitoring and mapping in precision agriculture applications. In this study, we developed a strategy for inter- and intra-row weed detection in early season maize fields from aerial visual imagery. More specifically, the Hough transform algorithm (HT) was applied to the orthomosaicked images for inter-row weed detection. A semi-automatic Object-Based Image Analysis (OBIA) procedure was developed with Random Forests (RF) combined with feature selection techniques to classify soil, weeds and maize. Furthermore, the two binary weed masks generated from HT and OBIA were fused for accurate binary weed image. The developed RF classifier was evaluated by 5-fold cross validation, and it obtained an overall accuracy of 0.945, and Kappa value of 0.912. Finally, the relationship of detected weeds and their ground truth densities was quantified by a fitted linear model with a coefficient of determination of 0.895 and a root mean square error of 0.026. Besides, the importance of input features was evaluated, and it was found that the ratio of vegetation length and width was the most significant feature for the classification model. Overall, our approach can yield a satisfactory weed map, and we expect that the obtained accurate and timely weed map from UAV imagery will be applicable to realize site-specific weed management (SSWM) in early season crop fields for reducing spraying non-selective herbicides and costs.

Research on farmland crop classification based on UAV multispectral remote sensing images

Abstract: UAV remote sensing technology is used to accurately and efficiently identify crops in farmland, providing technical support for obtaining crop category information in time and realizing precision agriculture. Obtain multi-spectral images of farmland peach trees, corn, weeds and other ground objects with quadrotor equipped with a multi-spectral camera, and use the method of principal component analysis to transform the multi-spectral image, retain the first 2 spectral bands with the most information content, and calculate 16 textures from 7 vegetation indices, 6 reflectances and 2 principal component bands, totaling 29 The item feature is used as the feature parameter of the classification sample. The ReliefF algorithm is used to screen the spectral features and texture features respectively. The 11 features obtained (blue reflectivity, green reflectivity, red edge reflectivity, normalized green band difference vegetation index, enhanced vegetation index, ratio vegetation index, first One principal component correlation, second principal component synergy, information entropy, second moment, correlation). from the screening form the A-group feature data set. All the spectral features extracted from the image form the B-group feature data set and all the extracted texture features form the C-group feature data Set, all 29 features form the D feature data set. SVM support vector machine is used to supervise and classify the 4 groups of feature data sets, and the accuracy of the classification results obtained is evaluated. The accuracy of the supervised classification model trained on the single category feature data sets of B and C is poor, and the SVM supervised classification model trained on the feature data set of the ReliefF algorithm has a better classification effect, with an overall accuracy of 90.09%, and the Kappa coefficient of 0.86. The SVM model trained on the feature data set D has the best classification effect, with an overall accuracy of 92.01%, and the Kappa coefficient of 0.89. The ground object classification based on UAV multi-spectral images is efficient and feasible, which provides a reference for timely acquisition of crop planting structure in farmland. Keywords: UAV, Ground object identification, Multispectral remote sensing, ReliefF algorithm DOI: 10.33440/j.ijpaa.20210401.153 Citation: Yang D J, Zhao J, Lan Y B, Wen Y T, Pan F J, Cao D L, Hu C X, Guo J K. Research on farmland crop classification based on UAV multispectral remote sensing images. Int J Precis Agric Aviat, 2021; 4(1): 29–35.

RoboWeedSupport-Semi-Automated Unmanned Aerial System for Cost Efficient High Resolution in Sub-Millimeter Scale Acquisition of Weed Images

Recent advances in the Unmanned Aerial System (UAS) safety and perception systems enable safe low altitude autonomous terrain following flights recently demonstrated by the consumer DJI Mavic PRO and Phamtom 4 Pro drones. This paper presents the first prototype system utilizing this functionality in form of semi-automated UAS based collection of crop/weed images where the embedded perception system ensures a significantly safer and faster gathering of weed images with sub-millimeter resolution. The system is to be used when the weeds are at cotyledon stage and prior to the harvest recognizing the grass weed species, which cannot be discriminated at the cotyledon stage.

A Comparative Analysis of Weed Images Classification Approaches in Vegetables Crops

A Comparative Analysis of Weed Images Classification Approaches in Vegetables Crops

Smartphone-based hierarchical crowdsourcing for weed identification

Weed infestation is a common problem in agriculture that adversely affects crop production. Given severe constraints on the budget of many land-grant universities due to the economic downturn, extension services or agencies responsible for educating farmers and assisting them with the application of advancements in agricultural research, have taken a hit. To adapt to the current economic climate without adversely affecting the quality of programs for weed management, we present a hierarchical system that uses images captured using a smartphone application, a backend image processing algorithm, and two levels of crowdsourcing to identify weed images. The first of the two crowdsourcing levels consist of a non-expert crowd contributed by Amazon Mechanical Turk (AMT) and the second level consists of a crowd composed of experts such as county extension agents. We present a probabilistic decision engine to determine the suitability of two levels of crowdsourcing for identifying the weed image. We have evaluated the designed system using test weed images and we show that 80% of the weeds in our test set can be identified using the low cost AMT crowd while incurring a maximum latency of 3h. Our system can help reduce the loses caused by the delay in identifying weeds, and hence, lead to quick remedial control practices applied to contain weed infestations.

‘Weed it out’a Digital Reference Tool for Agricultural Community

India is a rural country with most of its population engaged in agriculture, which is marred by many problems like inclement weather, poor fertile soils, pests and weeds. Weeds cause extensive damage to crops and are often mismanaged due to improper information. The study is an attempt to build a digital collection of Indian weeds using greenstone digital library software. Extensive field trips were made to capture images of weeds with high resolution cameras. Specimens of weeds spotted in desiccated state or not spotted in the fields due to short life span were taken from authentic secondary sources. Each specimen was assigned adequate metadata for easy identification and reference. Through the use of greenstone open source software, a digital library was built. The digital collection is fully text searchable and browsable through scientific name, habitat, season, local name and family. The collection can serve as a digital reference tool for farmers, students, researchers for identification of different weeds.

Research of Image Segmentation Based on Iterative Threshold

Image segmentation is the key step in image recognition,the result of segmentation affects the one of recognition directly.The article introduces the concept and detailed definition of the image segmentation. The segmentation algorithm of iterative threshold in detail. According to the intrinsic characteristics of weed images, just can use the iteration threshold segmentation method, and implements them by Matlab programme, then processes three weed images, respectively to obtain effective results , and establishes a good base for the pick-up of the target character.

Weed Image Classification using Wavelet Transform, Stepwise Linear Discriminant Analysis, and Support Vector Machines for an Automatic Spray Control System *

We tested and validated the accuracy of wavelet transform along with stepwise linear discriminant analysis (SWLDA) and support vector machines (SVMs) for crop/weed classification for real time selective herbicides systems. Unlike previous systems, the proposed algorithm involves a pre-processing step, which helps to eliminate lighting effects to ensure high accuracy in real-life scenarios. We tested a large group of wavelets (46) and decomposed them up to four levels to classify weed images into weeds with broad leaves versus weeds with narrow leaves classes. SWLDA was then employed to reduce the feature space by extracting only the most meaningful features. Finally, the features provided by SWLDA were fed to the SVMs for classification. The proposed method was tested on a database of 1200 samples, which is a much larger database size than that studied previously (200-400 samples). Using confusion matrices, the crop/ weed classification results obtained using different wavelets at different decomposition levels were compared, and this approach was also compared with existing techniques that use statistical and structural approaches. The overall classification accuracy obtained using the symlet wavelet family was 98.1%. These results represent an improvement of 14% in performance compared with existing techniques.

Microcontroller based roller contact type herbicide applicator for weed control under row crops

A microcontroller based manually operated three row roller contact type herbicide applicator was designed and developed for control of the weed population in field crops. A control system was developed to apply the quantity of the herbicide based on quantified weed information. The unit consists of a camera for capturing the images of weeds, MATLAB software for image acquisition and processing in a laptop, a serial port communication for communicating between laptop and controller, a microcontroller for controlling the application of herbicide through a relay, and a dc solenoid valve for variable rate application of herbicide on the applicating roller. The captured image was analyzed by the image processing toolbox in the MATLAB software, to extract weed information in the image, which is then transferred to a microcontroller using serial port interface. The microcontroller activates the solenoid valve using a relay according to algorithm for decision of the herbicide amount. Field test results of the machine indicate an average of 50% saving in the amount of herbicide, with weeding efficiency of 90%.