Sugarcane node recognition technology based on wavelet analysis

In order to realize automatic production of sugarcane seed, a sugarcane node recognition analysis based on multi-threshold and multi-scale wavelet transform is proposed. The laser displacement sensor is used to obtain the surface contour signal of sugarcane. After analyzing the signal characteristics, the 5th order of Daubechies wavelet base db5, is chosen as the wavelet mother function, and the signal is decomposed to eight layers by discrete wavelet. The fifth, sixth, and seventh layer coefficients are extracted to perform threshold processing, and the reconstructed signals processed by each threshold value are superimposed to characterize the characteristics of the sugarcane nodes. A multisensory redundancy algorithm based on Gauss membership function is proposed to improve the accuracy of recognition. Experiments show that the recognition rate of the algorithm is 100%, the maximum positioning error is less than 2.5 mm, and the maximum delay is 0.25 s. Compared with other four algorithms based on image processing, the proposed algorithm has higher effectiveness and recognition rate.