Abstract:Wisdom agriculture is a significant stage goal in the process of agricultural modernization development. Wisdom agriculture promotes the integration of agricultural informatization and intelligence. In recent years, the new models of intelligent agriculture based on artificial intelligence has developed rapidly. In this paper, 3D laser point cloud is used as research data to carry out in-depth research in the field of agriculture based on deep learning technology and point cloud. In this study, the deep learning model Pointnet ++ was used to segment the rapeseed point cloud data in the field: (1) The color enhancement algorithm of HSV color space was used to achieve color threshold segmentation of rapeseed crop point cloud data in complex field environment, and Statistical Outlier Filter and Super-Voxel Clustering were used to segment group rapeseed point cloud respectively. Finally, two groups of pure rapeseed point cloud data were obtained. (2) In this research, six original rapeseed point cloud data sets were used as datasets to train and test the segmentation performance of Pointnet++ (Multi-scale Grouping, MSG) deep learning model for rapeseed point cloud. Intersection over Union(IoU) was taken as the evaluation index of point cloud segmentation accuracy. The IoU of rape point cloud data processed by the three segmentation methods were 0.7748, 0.8019 and 0.8260, respectively. The results show that the segmentation performance of the deep learning model based on Pointnet ++ (MSG) is higher than that of the conventional point cloud segmentation algorithm. Compared with the conventional point cloud segmentation models, the point cloud segmentation based on deep learning framework shows better performance. The construction of a deep learning framework for crop point cloud segmentation and classification in the field requires the corresponding feature extraction processing based on the geometric structure or attributes of specific crops. In the context of the rapid development of agricultural big data, the deep learning framework in the field of agriculture is robust to deal with complex field environment, and the application of deep learning to agricultural research has a good prospect.

Segmentation of Green Vegetation in the Field Using Deep Neural Networks*

Deep learning implementation of image segmentation in agricultural applications: a comprehensive review

PD-SegNet: Semantic Segmentation of Small Agricultural Targets in Complex Environments

Adaptive segmentation of field image for green plants

A Semantic Segmentation Approach Based On Deeplab Network In High-Resolution Remote Sensing Images

A Novel Denoising Autoencoder Assisted Segmentation Algorithm for Cotton Field

Farmland Segmentation from Remote Sensing Images Using Deep Learning Methods

Sementing the Field of Rapeseed from 3D Laser Point Cloud Using Deep Learning

Fully Convolutional Network for Rice Seedling and Weed Image Segmentation at the Seedling Stage in Paddy Fields.

A Point-Cloud Segmentation Network Based on SqueezeNet and Time Series for Plants

Plant-part Segmentation Using Deep Learning and Multi-View Vision

A novel green apple segmentation algorithm based on ensemble U-Net under complex orchard environment

Bagging Improves the Performance of Deep Learning-Based Semantic Segmentation with Limited Labeled Images: A Case Study of Crop Segmentation for High-Throughput Plant Phenotyping

Deep semantic segmentation of unmanned aerial vehicle remote sensing images based on fully convolutional neural network

Semantic Segmentation of Crop and Weed using an Encoder-Decoder Network and Image Enhancement Method under Uncontrolled Outdoor Illumination

Transferring learned patterns from ground-based field imagery to predict UAV-based imagery for crop and weed semantic segmentation in precision crop farming

Greenery Segmentation In Urban Images By Deep Learning

Smallholder Crop Area Mapped with a Semantic Segmentation Deep Learning Method.

Efficient Semantic Segmentation for Large-Scale Agricultural Nursery Managements via Point Cloud-Based Neural Network

Extraction of Agricultural Fields via DASFNet with Dual Attention Mechanism and Multi-scale Feature Fusion in South Xinjiang, China

Semantic Segmentation of Vegetation in Remote Sensing Imagery Using Deep Learning