Abstract:With the development of digital image technology, we can easily obtain a large number of crop growth images. Through effective analysis of the image, the growth information of crops can be obtained, which can better direct agricultural production. The efficiency of traditional seedling growth monitoring is low, especially in large-scale farmland, which takes a lot of time. Artificial method timely restricts scientific decision-making of cultivation crops. The progress of machine vision and image processing technology provides a new way for harmlessly monitoring of crop seedling growth .The results of image analysis can help agricultural producers to understand the growth of crop seedlings quickly and accurately, so as to take effective management as soon as possible. In this paper, the images of sunflower seedling collected in farmland environment are taken as the research object. The main research content is to segment green crops from soil background. Segmentation method of sunflower seedling image based on color features and Ostu threshold segmentation is proposed. The method is simple in calculation, and can adapt to the segmentation of farmland environment images, which lays the foundation for crop recognition process. Based on the image recognition results, the algorithm locates the seedlings. Through the rapid identification of sunflower seedlings, it is possible to fill the gaps with seedlings where the seedlings are less distributed. On the contrary, if the seedlings are too dense, the number of seedlings needs to be reduced. The algorithm provides a basis for precise management. The results show that the algorithm with extra green feature can quickly and effectively identify sunflower seedlings from background, and locate the seedlings based on the image recognition results. This algorithm is not sensitive to soil moisture and light conditions, and is less affected by crop residual coverage, so it can adapt to different soil environment which realize the non-destructive monitoring of sunflower seedlings.

Research on Plant Seeds Recognition Based on Fine-Grained Image Classification

A dataset for fine-grained seed recognition

Research on the Fine-Grained Plant Image Classification

Rapidly and exactly determining postharvest dry soybean seed quality based on machine vision technology.

AIseed: an Automated Image Analysis Software for High-Throughput Phenotyping and Quality Non-Destructive Testing of Individual Plant Seeds

On the Efficacy of Handcrafted and Deep Features for Seed Image Classification

Multi-Feature Recognition of Healthy Vegetable Seedlings Based on Machine Vision Technology

Recognition and location of crop seedlings based on image processing

Study of Recognition Method of Leguminous Weed Seeds Image

Approach for identifying crop seeds with similar appearances using hyperspectral images and improved ResNet 18 based on cloud platform

High-throughput soybean seeds phenotyping with convolutional neural networks and transfer learning

Research on variety identification of common bean seeds based on hyperspectral and deep learning

Maize seed appearance quality assessment based on improved Inception-ResNet

Rapid and Accurate Varieties Identification of Different Crop Seeds Under Sample-Limited Condition Based on Hyperspectral Imaging and Deep Transfer Learning

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

Classification of Hybrid Seeds Using Near-Infrared Hyperspectral Imaging Technology Combined with Deep Learning

Classification of Weed Seeds Based on Visual Images and Deep Learning

A maize seed variety identification method based on improving deep residual convolutional network

Research on terahertz image analysis of thin-shell seeds based on semantic segmentation

High-Throughput Phenotyping of Seed Quality Traits Using Imaging and Deep Learning in Dry Pea

Rapid and Accurate Varieties Classification of Different Crop Seeds Under Sample-Limited Condition Based on Hyperspectral Imaging and Deep Transfer Learning