Abstract
The leaf is the organ that is crucial for photosynthesis and the production of nutrients in plants; as such, the number of leaves is one of the key indicators with which to describe the development and growth of a canopy. The irregular shape and distribution of the blades, as well as the effect of natural light, make the segmentation and detection process of the blades difficult. The inaccurate acquisition of plant phenotypic parameters may affect the subsequent judgment of crop growth status and crop yield. To address the challenge in counting dense and overlapped plant leaves under natural environments, we proposed an improved deep-learning-based object detection algorithm by merging a space-to-depth module, a Convolutional Block Attention Module (CBAM) and Atrous Spatial Pyramid Pooling (ASPP) into the network, and applying the smoothL1 function to improve the loss function of object prediction. We evaluated our method on images of five different plant species collected under indoor and outdoor environments. The experimental results demonstrated that our algorithm which counts dense leaves improved average detection accuracy of 85% to 96%. Our algorithm also showed better performance in both detection accuracy and time consumption compared to other state-of-the-art object detection algorithms.
Highlights
The phenotype of plant refers to all observable characteristics of a plant, including its physical morphology as well as biochemical and physiological properties
The segmentation-based object detection algorithms for leaf counting employed by previous studies have relatively high false rates and cannot provide clear segmentations for dense and overlapped leaves
Here we developed the improved CenterNet, i.e., a key-point-based onestage object detection deep learning detector
Summary
The phenotype of plant refers to all observable characteristics of a plant, including its physical morphology as well as biochemical and physiological properties. Vukadinovic and Polder [14] proposed a method of combining a supervised classification and a simple artificial neural network to segment the plant regions, and using watershed transformation to identify individual leaves. This approach uses ground truth images to mask the plant and background pixels, and is not suitable for automatic processing. Artificial intelligence methods such as convolutional neural networks (CNNs) have made important progress in plant phenotyping, and have brought wide applications in plant classification, fruit detection and leaf segmentation [19].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
