Edge AI Insect Monitoring and Alarm in Litchi Farm
Hardware: reComputer J1020v2 with NVIDIA Jetson Nano
Application: Insect Detection
Industry: Smart Farming
Deployment Location: China
Conopomorpha sinensis infestation is a very important factor affecting litchi harvest. The goal of this solution is to help farmers quickly locate and identify the root borer on the branches, leaves, and fruits of litchi trees, and apply pesticides at designated points, to finally achieve timely pest control in the early stage of risk. At the same time, it can also avoid excessive use of pesticides for environmental protection and cost saving.
Challenge
Detecting insects on Lichi trees poses significant challenges due to the small size of each individual insect, making accurate detection and segmentation exceedingly difficult. The inherent difficulty lies in precisely identifying and segmenting insects, especially when they are occluded or overlapping. The small scale of these insects adds complexity to the detection process.
Furthermore, the image dataset collected from the field introduces additional issues, as varying illumination conditions and shades across different images can adversely impact detection accuracy. These environmental factors create a dynamic and challenging landscape for insect detection deployment, necessitating robust techniques to address the intricacies showed by the size variability and environmental conditions in the dataset.
Solution
The object detection model can be successfully deployed on Jetson Nano for inspects, trained based on CNN darknet53 architecture and COCO dataset. To minimize the overlap of bounding boxes caused by several identified pests gathered and overlapped together, so that the predicted number of pests is close to the actual number in the image, the bounding boxes of adjacent clusters should be merged, and remove the minimal clusters.
Jetson Nano can easily pre-process the huge amount of input images and send the decision signal to the electrical spray module for opening/closing control based on the results of object detection. As shown from the detection results it delivers good robustness to complex scenes even with light occlusion or dark shadow background.
Seeed NVIDIA Jetson Ecosystem
Seeed is an Elite partner for edge AI in the NVIDIA Partner Network. Explore more carrier boards, full system devices, customization services, use cases, and developer tools on Seeed’s NVIDIA Jetson ecosystem page.
Join the forefront of AI innovation with us! Harness the power of cutting-edge hardware and technology to revolutionize the deployment of machine learning in the real world across industries. Be a part of our mission to provide developers and enterprises with the best ML solutions available. Check out our success case study catalog to discover more edge AI possibilities!
Discover infinite computer vision application possibilities through our vision AI resource hub!
Take the first step and send us an email at [email protected] to become a part of this exciting journey!
Download our latest Jetson product Catalog to find one option that suits you well. If you can’t find the off-the-shelf Jetson hardware solution for your needs, please check out our customization services, and submit a new product inquiry to us at [email protected] for evaluation.
About Author
