RT-DETR-ME: a Lightweight Corn and Weed Detection Algorithm Based on Improved RT-DETR

Weed detection in multi-spectral images of cotton fields

Weed is one of the main reason for getting less production in agriculture field. At present, farmers are using herbicides to control the weed but it's having negative impact on crop production. To increase the crop production, farmers want to reduce the usage of herbicides. One of the machine learning algorithm, convolution neural network used to classify the weed and crop with high accuracy. CNN gives the best performance in some critical situations also, such as collected images of different lighting conditions, identification of plant species, overlapped crop with weed and designing an autonomous patch sprayers. This review article presents a brief overview of some significant research efforts in weed detection system by using image processing techniques and convolution neural network in machine learning algorithms. Not only CNN, this article reviewed different successful machine learning techniques of supervised and unsupervised learning approaches such as support vector machines, random forest and artificial neural network. This paper aims at providing different challenges and imperatives of various convolution neural network algorithms used in weed detection. Finally, compared different convolution neural network architectures and finds the best CNN architecture used for weed identification system with respect to accuracy.

Weed identification techniques that are effective can lower the cost of weed control while also improving crop quality and output. Herbicides are currently the most frequent method of weed management. In order to deal with the issues of weed man-agement in agriculture, accurate crop and weed discrimination is essential. This work deals with the performance comparison of different classification algorithms namely Convolutional Neural Network(CNN), Graph Convolutional Networks(GCN), k-Nearest Neighbor(k-NN), and Support Vector Machine(SVM). Feature extraction algorithms, Local Directional Pattern(LDP) and Local Directional Relation Pattern(LDRP) are used alongside SVM and kNN. For data expansion, Generative Adversarial Network(GAN) was used, which generates synthetic samples of images from the dataset. The result of the study shows that CNN gives the maximum accuracy of 90.58% in ‘Crop and weed detection’ dataset when compared with other classification models used for investigation in our paper.

The spread of aquatic invasive plants is a major concern in several zones of the world's geography. These plants, which are not part of the natural ecosystem, cause a negative impact to the environment, as well as to economy and society. In Spain, large areas of Guadiana (the second longest river in Spain) have been invaded by such plants. Among the strategies to address this problem, monitoring and detection play an important role to control the spatio-temporal distribution of the invasive plants. The main objective of this work is to develop a methodology able to automatically detect the geo-location of aquatic invasive plants using remote sensing and machine learning techniques. To this end, several classification algorithms have been applied to freely available multispectral satellite imagery, collected by ESA's Sentinel-2 satellite. A quantitative and comparative assessment is conducted using different machine and deep learning algorithms, from classical methods such as unsupervised K-means to supervised random forests (RFs) and convolutional neural networks (CNNs). This study also proposes a methodology for validating the obtained classification results, generating synthetic ground truth images based on available high spatial resolution imagery. The obtained results demonstrate the suitability of some of the considered algorithms for automatic detection of aquatic weeds in satellite images with medium spatial resolution.

Due to the existence of cotton weeds in a complex cotton field environment with many different species, dense distribution, partial occlusion, and small target phenomena, the use of the YOLO algorithm is prone to problems such as low detection accuracy, serious misdetection, etc. In this study, we propose a YOLOv8-DMAS model for the detection of cotton weeds in complex environments based on the YOLOv8 detection algorithm. To enhance the ability of the model to capture multi-scale features of different weeds, all the BottleNeck are replaced by the Dilation-wise Residual Module (DWR) in the C2f network, and the Multi-Scale module (MSBlock) is added in the last layer of the backbone. Additionally, a small-target detection layer is added to the head structure to avoid the omission of small-target weed detection, and the Adaptively Spatial Feature Fusion mechanism (ASFF) is used to improve the detection head to solve the spatial inconsistency problem of feature fusion. Finally, the original Non-maximum suppression (NMS) method is replaced with SoftNMS to improve the accuracy under dense weed detection. In comparison to YOLO v8s, the experimental results show that the improved YOLOv8-DMAS improves accuracy, recall, mAP0.5, and mAP0.5:0.95 by 1.7%, 3.8%, 2.1%, and 3.7%, respectively. Furthermore, compared to the mature target detection algorithms YOLOv5s, YOLOv7, and SSD, it improves 4.8%, 4.5%, and 5.9% on mAP0.5:0.95, respectively. The results show that the improved model could accurately detect cotton weeds in complex field environments in real time and provide technical support for intelligent weeding research.

Unmanned Aerial Vehicles (UAVs) have become an integral part of several real-world applications. Their combination with other evolving paradigms such as image recognition using machine learning or deep learning algorithms has contributed to the suitability for use in smart agriculture and weed detection applications. In this paper, the performances of the Artificial Neural Network (ANN) and AlexNet algorithms for weed detection using UAV-based images have been studied. An image dataset containing 15336 segments with the following breakdown: 3249 of soil, 7376 of soybean, 3520 grass and 1191 of broadleaf weeds has been used. The partitioning for train and test sets has been done in the ratio of 70:30. Simulation results show that the conventional ANN algorithm provide an accuracy of 48.09% while the AlexNet algorithms gives an accuracy 99.8% on the test dataset.

Robotic plant-specific spraying can reduce herbicide usage in agriculture while minimizing labor costs and maximizing yield. Weed detection is a crucial step in automated weeding. Currently, weed detection algorithms are always evaluated at the image level, using conventional image metrics. However, these metrics do not consider the full pipeline connecting image acquisition to the site-specific operation of the spraying nozzles, which is vital for an accurate evaluation of the system. Therefore, we propose a novel application-specific image-evaluation method, which analyses the weed detections on the plant level and in the light of the spraying decision made by the robot. In this paper, a spraying robot is evaluated on three levels: (1) On image-level, using conventional image metrics, (2) on application-level, using our novel application-specific image-evaluation method, and (3) on field level, in which the weed-detection algorithm is implemented on an autonomous spraying robot and tested in the field. On image level, our detection system achieved a recall of 57% and a precision of 84%, which is a lower performance than detection systems reported in literature. However, integrated on an autonomous volunteer-potato sprayer-system we outperformed the state-of-the-art, effectively controlling 96% of the weeds while terminating only 3% of the crops. Using the application-level evaluation, an accurate indication of the field performance of the weed-detection algorithm prior to the field test was given and the type of errors produced by the spraying system was correctly predicted.

SP-DETR: Superior point weak semi-supervised DETR with teacher–student paradigm for crop and weed detection

Controlled comparison of machine vision algorithms for Rumex and Urtica detection in grassland

An improved algorithm for sesame seedling and weed detection based on YOLOv7

Digital images of the soybean and three species of weeds with different backgrounds were acquired for image processing. An adaptive neuro-fuzzy inference system (ANFIS) from MATLAB® 6.1 was used to obtain a preliminary separation of plant and background. To train the fuzzy inference system (FIS), red-green-blue (RGB) values were first converted to HSV color space (hue, saturation, value), and samples from plant and background were visually selected from the images to be used as inputs. These were then mapped to a binary output that indicated the presence or absence of plant. Subtractive clustering was used to generate the FIS with a total of 100 epochs. Three clusters, and therefore three rules, were found by the ANFIS. This resulted in a total of 9 membership functions. Source code for a genetic algorithm was written using the toolbox of MATLAB® 6.1 to adjust the membership functions to reduce misclassification and improve segmentation. The membership function parameters were used to generate chromosome genes. The initial population of 50 chromosomes was generated using 30% noise perturbation of the membership functions originally given by ANFIS. A final segmentation of background and plant was achieved after 10 generations. Between evaluations crossover and mutations were applied. The results showed that the adaptive neuro-fuzzy inference system and genetic algorithm segmentation system was capable of eliminating areas that were misclassified as plant and also generally the final segmentation.

Autonomous robotic weeding systems in precision farming have demonstrated their full potential to alleviate the current dependency on herbicides or pesticides by introducing selective spraying or mechanical weed removal modules, thus reducing the environmental pollution and improving the sustainability. However, most previous works require fast weed detection system to achieve real-time treatment. In this paper, a novel computer vision based weeding control system is presented, where a non-overlapping multi-camera system is introduced to compensate the indeterminate classification delays, thus allowing for more complicated and advanced detection algorithms, e.g. deep learning based methods. The suitable tracking and control strategies are developed to achieve accurate and robust in-row weed treatment, and the performance of the proposed system is evaluated in different terrain conditions in the presence of various delays.

Precision agriculture, which uses deep learning and artificial intelligence (AI) to enable real-time crop monitoring and management, has completely changed contemporary farming. This study presents a deep learningbased weed detection algorithm that classifies agricultural imagery into crop and weed categories using the InceptionV3 architecture. To improve model dependability, the Kaggle dataset was restructured into a structured directory format and divided into subsets for training (70%) validation (15%) and testing <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(15 \%)$</tex>. Training was conducted using a binary classification method with a sigmoid activation function, yielding an astounding 98.46% accuracy on the test set. GradCAM visualization was used to pinpoint important areas impacting grouping results in order to put light on model predictions. Grad-CAM visualization was used to identify important areas impacting categorization results in order to shed light on model predictions. The model is able to effectively identify weeds and crops by using the main morphological, textual, and color components to assist precision farming by using drones or robotic sprayers to apply the necessary herbicides. The paper provides a foundation to real-time surveillance, and the next round of investigation will be to implement it in the field and add more diseases and pests to the list.

A lightweight weed detection algorithm designed based on YOLOv9

Weed control is essential in agricultural productivity as weeds act as a pest to crops. The conventional methods of weed removal are time-consuming and require more manual labour work. Hence there is a need to automate this process. The objective of the proposed system is to detect weed from crop using machine learning algorithms. The exhaustive dataset is collected for four different commercial crops and two types of weeds such as Para grass and Nutsedge. Excess green method and Otsu’s thresholding is used for masking the soil and extract the region of interest. The shape features of an image are extracted to provide distinguish properties between weed and crop. The classification of weed and crop has experimented with three different classifiers: Support Vector Machine, Artificial Neural Network and Convolutional Neural Network. The performance comparison of weed detection algorithms is executed on the Open CV and Keras platform using python language.