System for detecting fires in live images captured by aerial vehicles, has cross-stage partial network provided with CSP darknet as backbone which addresses repeated gradient information and integrates gradient changes into feature map
	ETHIRAJULU V; SAI M V N; GOPICHAND J; AKHIL J; SUMITH K S; JANISHAM
	2023-10-16
专利权人	BHARATH HIGHER EDUCATION & RES INST (BHAR-Non-standard)
申请日期	2023-10-16
专利号	IN202341069639-A
成果简介	NOVELTY - The system has a cross-stage partial network provided with a CSP darknet as backbone which addresses repeated gradient information and integrates gradient changes into the feature map to reduce parameters and computational load. A PANet is used as the neck in Yolov5 to enhance information flow while Yolo acts as the head of the network and generates feature maps of different sizes for multi-scale prediction. An efficient net utilizes efficient net as backbone to enable the learning of complex features and employ a bi-directional pyramid network as its neck to facilitate easy information flow. USE - System for detecting fires in live images captured by aerial vehicles. Uses include but are not limited to BowFire, FD-dataset, Forestry Images, VisiFire (10,581 images, with2976fireimages and 7605 non-fire images). Yolov5 incorporates a cross-stage partial network withCSP darknet as its backbone. ADVANTAGE - The system trains deep neural networks with large amounts of data, and allows the network to automatically learn and improve from experience.
IPC 分类号	A62C-003/02 ; B64C-039/02 ; G06K-009/62 ; G06N-003/04 ; G06N-003/08
国家	印度
专业领域	信息技术
语种	英语
成果类型	专利
文献类型	科技成果
条目标识符	http://119.78.100.226:8889/handle/3KE4DYBR/19895
专题	中国科学院新疆生态与地理研究所
作者单位	BHARATH HIGHER EDUCATION & RES INST (BHAR-Non-standard)
推荐引用方式 GB/T 7714	ETHIRAJULU V,SAI M V N,GOPICHAND J,et al. System for detecting fires in live images captured by aerial vehicles, has cross-stage partial network provided with CSP darknet as backbone which addresses repeated gradient information and integrates gradient changes into feature map. IN202341069639-A[P]. 2023.

条目包含的文件
条目无相关文件。

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。