Lightning has been one of the most talked-about natural disasters worldwide in recent years, as it poses a great threat to all industries and can cause huge economic losses. Thunderstorms are often accompanied by natural phenomena such as lightning strikes and lightning, and many scholars have studied deeply the regulations of thunderstorm generation, movement and dissipation to reduce the risk of lightning damage. Most of the current methods for studying thunderstorms focus on using more complex algorithms based on radar or lightning data, which increases the computational burden and reduces the computational efficiency to some extent. This paper proposes a raster-based DWT (discrete wavelet transform) method for thunderstorm identification, this method uses DWT, CFSFD (clustering algorithm for fast search and finding density peaks) algorithm and ADTD (active divectory topology diagrammer) lightning location data for thunderstorm identification. The advantage of this method is that it supports different spatial resolutions and can identify any shape and number of thunderstorms at the same time and in the same area. It is effective in eliminating some of cluttered, scattered lightning data and extracting dense areas of thunderstorms. Furthermore, the method has a time complexity of O(n), and the computational efficiency is significantly better than the current TITAN (thunderstorm identification, tracking, analysis, and nowcasting) algorithm, which provides a good basis for subsequent extrapolation studies of thunderstorms.

In the work, the concept of orthogonal vector-valued trivariate wavelet packets, which is a generalization of uniwavelet packets, is introduced. A new method for constructing them is developed, and their characteristics is discussed by using time-frequency analysis method, matrix theory and finite group theory. Orthogonality formulas are established.

The notion of matrix-valued multiresolution analysis. A procedure for designing orthogonal matrix-valued univariate wavelet packets is presented and their orthogonality properties are discussed by means of time-frequency analysis method, matrix theory and functional analysis method. Three orthogonality formulas concerning these wavelet packets are obtained. Finally, one new orthonormal basis of are obtained by constructing a series of subspaces of orthogonal matrix-valued wavelet packets.