Study On Data Acquisition Of Ecological Factors And Processing Algorithms For Ginseng Cultivation

Ginseng may naturally enter a"period of dormancy"as it grows.This period could be one or two months,one or two years,or even longer,depending on the habitat factors.In harsh weather,ginseng may activate a self-protection mechanism and go dormant if the ecological habitat factors are not conducive to ginseng growth.Therefore,researchers and relevant practitioners are constantly faced with the challenge of ensuring the sustainable growth of ginseng.It is proposed to install many wireless sensors for three-dimensional ginseng cultivation in a sustainable greenhouse substrate to regulate the ecological habitat factors and improve ginseng cultivation efficiency and sustainability,making an intelligently controllable habitat for precise cultivation.Because of the massive hyper-spectral image characteristic data capturing wavelengths of biological factors,compression becomes critical in this field of study.Furthermore,the wireless sensor network’s life cycle,the reliable transmission of biological factors,and the mining of characteristic data have become research priorities in order to achieve long-term ginseng growth.Therefore,the following materials were used in this study to efficiently optimize the ginseng cultivation mode,substrate condition,and cultivation method,as well as to improve the precision of the quantized algorithm for data acquisition and processing of the biological factors in ginseng cultivation:The first aspect focuses on the lossless compression of ultra-spectral image data of biological factors for ginseng cultivation.An inter-spectral redundancy removal quantization algorithm based on invertible matrix transformations,three-band parametric spectral integer reversible transformation,was proposed to reduce data storage complexity,data transmission time,and processing time of the hyper-spectral image during ginseng cultivation and further improve transmission and storage efficiencies.Then,to remove spatial redundancy,the discrete wavelet transform（DWT）was used,and the SPIHT algorithm was used for coding to find the optimal number of dimensions of the reversible transformation matrix.In addition,the integral matrix reversible transformation-based redundancy removal was investigated to optimize bit allocation and determine parameters.In comparison to traditional lossless compression algorithms,the proposed algorithm’s lossless compression ratio of each band changed rapidly with band change and increased significantly on average.This algorithm is useful for researching hyper-spectral image sensors in agricultural engineering and is simple to implement in hardware.The second is about the life cycle of a wireless sensor network related to ginseng cultivation.There are numerous issues with ginseng cultivation,including highly dispersed data sampling points of ginseng cultivation-related biological factors,uneven geographical distribution,power supply difficulties,and the limited life expectancy of wireless sensor network nodes.Therefore,to extend the nodes’expected life and improve the network’s life cycle,a ginseng cultivation-oriented efficient and energy-saving wireless sensor network algorithm was proposed.The communication ranges of sensors in the wireless network were adjusted based on the characteristics of wireless sensors and the advantage of the particle swarm optimization（PSO）algorithm in optimization,and positive,high-energy nodes were selected as cluster-head nodes or data forwarding nodes.In comparison to commonly used algorithms,the number of wireless network nodes,host computer position,and distance monitoring range of the center node as influential factors of communication distance had a significant impact on the network’s life cycle.The life cycle of the network based on the proposed algorithm gradually increased,with large-scale node failure appearing at the end.The third relates to the network’s dependable data transmission.There are problems in the habitats assigned to different networks for ginseng cultivation,such as data package loss in wireless sensor node data transmission,long delays caused by an increase in the number of conflicts and retransmissions,and the impact of redundant information fusion efficiency on the package loss rate.A multilayer perceptron-based reliable data transmission algorithm for wireless sensor networks was proposed to address these issues.In detail,an RBF neural network acquired a reliable data transmission situation based on its adaptivity and highly parallel computing capability;the reliable data transmission was realized through retransmission and redundancy.In comparison to commonly used algorithms,the proposed algorithm has a long reliable transmission distance,a low package loss rate,a small transmission delay on average as the number of retransmissions increases,a high redundant information fusion efficiency of data,and a high increment in the average network reliability in different ginseng cultivation scenes.The fourth is concerned with the mining of biological factors for ginseng cultivation.Ginseng quality is closely related to the biological factors’characteristic data.The traditional k-means clustering algorithm,on the other hand,is highly dependent on the initial parameters.Meanwhile,the processing of biological factors for ginseng cultivation is susceptible to the interferences of noisy data and outlier data points and the mixed biological factors are the cause of a poor mining effect and single type of the mined data.In this regard,an algorithm for mining biological factor characteristic data for ginseng cultivation was proposed,which ranked abnormal factors in descending order.Based on the high-dimensional mixed attributes of the biological factors in ginseng cultivation,the k value was generated automatically.The globally optimal probability density was then obtained through coordinate rotation.The experimental results show that the proposed algorithm’s time efficiency was superior to that of typical algorithms during the iteration.The advantage of this algorithm in clustering mining was significant,which significantly reduced the mining error.Furthermore,the clustering results of the environment temperature and humidity,p H value of substrate,illumination,CO₂,and EC value of the nutrient solution for ginseng cultivation in a sustainable greenhouse were quantitatively analyzed,yielding the probability density of the biological factors’characteristic data.On this basis,a sustainable growth of ginseng can be realized by clustering quantitative regulation in the ginseng cultivation.