Research article

FORESTRY COMMERCIAL SIGNBOARD DESIGN AND MANAGEMENT STRATEGY BASED ON INTELLIGENT RECOGNITION OF STREET VIEW IMAGES

Wei Gan1 ,Lan Qiu1, Ziwei Zhang1*,Ying Lin1, Zongyu Hong1, Yuchen Lei1

Online First: January 08, 2023


With the development of computer technology and the popularization of various imaging devices, a large amount of landmark and street view image data has been accumulated on the Internet. Given a query image, how to efficiently and accurately retrieve images with similar content from these large-scale image sets has become an urgent need in many applications. Driven by the industrial revolution, the contemporary forestry industry has undergone tremendous changes in terms of scale, environment, and management methods. The visual design of the forestry industry has evolved from a monotonous signboard or signage to a brand image design with a unique personality. At the same time, there are various forms of media, such as digital media, print media, advertising media, etc. The fusion of various concepts and technologies makes the ancient art of visual design glow with new vitality. A forestry scene semantic segmentation model based on DeepalBV3+ is proposed, and an image enhancement algorithm based on generative confrontation network is designed to improve the semantic segmentation accuracy of forestry scenes under different lighting environments. The experimental results show that the OA, MA and MIoU evaluation indexes of the model are 0.9420, 0.7799 and 0.6925 respectively, which are significantly improved compared with the original model and can meet the requirements of forestry semantic segmentation.

Keywords

Intelligent recognition of street scenes; Forestry Commercial signboard design; Deep learningt