Pest24: A large-scale very small object data set of agricultural pests for multi-target detection

Wang, Qi-jin; Zhang, Shengyu; Dong, Shifeng; Zhang, Guangcai; Yang, Jui‐Chung; Li, Rui; Wang, Hongqiang

doi:10.1016/j.compag.2020.105585

Cited by 85 publications

(48 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A plethora of real-world applications is heavily dependent on accurate small object detection such as industrial (Qiu et al, 2019), agriculture (Wang et al, 2020;Zhang et al, 2020), sports (Xu et al, 2018), or security (Abe et al, 2008) applications. Small object detection is of particular interest for earth observation purposes such as the case of object detection in optical airborne (Wang et al, 2018), or satellite (Ren et al, 2018;Zhang et al, 2019) very high resolution data.…”

Section: Small Object Detectionmentioning

confidence: 99%

Scale-Awareness for More Accurate Object Detection Using Modified Single Shot Detectors

Tsironis¹,

Stentoumis²,

Lekkas³

et al. 2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Object detection performance is directly related to the apparent size of the object to be detected, thus most state-of-the-art algorithms dedicate different detection heads for each object size. In this work, we propose an end-to-end pipeline to adapt a single-shot object detector (SSD) to the underlying object size distribution of the target detection domain. Our contributions are the adjustments to the detector architecture and the introduction of a novel batch sampling method. To validate the effect of our method, we chose a task-specific highly specialized object detection and classification dataset of tomato fruits that apart from bounding box information, it also contains class information for three ripening stages of each tomato fruit.More specifically, the major motivation and contributions are discussed in relation to the recent bibliography. Next, an extensive analysis of our pipeline is presented, where the concept of scale alignment is thoroughly presented along with the novel sampling method. Following the results of a series of experiments, we conclude that our pipeline significantly improves over the “off-the-shelf” base single-shot detector and its detection performance is comparable to more elaborate algorithms, especially if we measure detection performance slightly disregarding box localization. Lastly, we include a stratified ablation study in the closing sections where we measure the impact of each step along our proposed SSD adaptation pipeline.

show abstract

Section: Small Object Detectionmentioning

confidence: 99%

Scale-Awareness for More Accurate Object Detection Using Modified Single Shot Detectors

Tsironis¹,

Stentoumis²,

Lekkas³

et al. 2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…"Wang, Q.J., Zhang, S.Y., Dong, S.F., Zhang, G.C., Yang, J., Li, R. and Wang, H.Q. ", in 2020 [22] studied that Accurate agriculture poses new challenges for onsite pest monitoring in real time based on the new AI technology generation. This paper establishes a large-scale standardised data collection of agricultural pests, called Pest24 to provide a large data resource for the training of profound learning models for the detection of pests.…”

Section: Pest24: a Large-scale Very Small Object Data Set Of Agricultural Pests For Multi-target Detectionmentioning

confidence: 99%

Recognition and Early Stage Detection of Phytophthora in a Crop Farm Using IoT

Vajpayee¹,

Yogi²

2021

Agro-Economic Risks of Phytophthora and an Effective Biocontrol Approach

View full text Add to dashboard Cite

Detection of agricultural plant pests is seen as one of the farmers’ problems. Automated Pest Detection Machine enables early detection of crop insects with advanced computer vision and image recognition. Innovative research in the field of agriculture has demonstrated a new direction by Internet of Things (IoT). IoT needs to be widely experienced at the early stage, so that it is widely used in different farming applications. It allows farmers increase their crop yield with reduced time and greater precision. For the past decade, climate change and precipitation have been unpredictable. Due to this, many Indian farmers are adopting smart methods for environment known as intelligent farming. Smart farming is an automated and IOT-based information technology (Internet of Things). In all wireless environments IOT is developing quickly and widely. The Internet of Things helps to monitor agricultural crops and thus quickly and effectively increase farmers’ income. This paper presents a literature review on IoT devices for recognizing and detecting insects in crop fields. Different types of framework/models are present which are explaining the procedure of insect detection.

show abstract

“…Wang et al (2020) [14] evaluated several detection methods for real-time monitoring of several pests attacking maize crops, using a dataset of 25.378 images of agricultural pests automatically captured by traps installed by the plantation. Among the methods chosen were YOLOv3 and Faster R-CNN that obtained accuracy in the act of detection of 63.54% and 51.72% respectively.…”

Section: Related Workmentioning

confidence: 99%

Automatic counting of cattle with Faster R-CNN on UAV images

Porto

Rezende

Astolfi³

et al. 2021

Anais Do XVII Workshop De Visão Computacional (WVC 2021)

View full text Add to dashboard Cite

It is remarkable the growth of the bovine herd in the last four decades however, the availability of areas for pasture did not follow the same trend and thus caused direct interference in the binomial quality and price of the final product. One of the ways to get around this interference is by the use of technologies to help minimize the handling costs, from the breeding in a controlled environment with the need of trained manpower in the confinement process. Thus, as opposed to the current format done manually and in restricted space, computer vision technology can mitigate the identification and counting of cattle problems using unmanned aerial vehicle (UAV). Attending to the objective outlined in this article demonstrates the use of the Faster R-CNN for counting cattle in feedlots employing aerial images, obtaining an average precision of 89.7% for the set of hyperparameters that differed most positively from the others in this experiment.

show abstract

Pest24: A large-scale very small object data set of agricultural pests for multi-target detection

Cited by 85 publications

References 12 publications

Scale-Awareness for More Accurate Object Detection Using Modified Single Shot Detectors

Scale-Awareness for More Accurate Object Detection Using Modified Single Shot Detectors

Recognition and Early Stage Detection of Phytophthora in a Crop Farm Using IoT

Automatic counting of cattle with Faster R-CNN on UAV images

Contact Info

Product

Resources

About