Robustness method used for voice recognition on mobile equipment during agricultural field data acquisition

Abstract

The invention provides a robustness method used for voice recognition on mobile equipment during agricultural field data acquisition. The robustness method is based on a noiseproof method which combines mobile equipment voice recognition characteristic compensation with model compensation, a non-stationary noise environment of a typical application scene is acquired according to the agricultural field data, steady noise-resistant voice characteristic parameters are searched, characteristic extracted from noise-containing voice is processed, and deviation, caused by noise, between the noise-containing voice characteristic and pure voice characteristic is removed, thus the accuracy rate of system recognition is effectively improved. The robustness method provided by the invention has low complexity and is easier to implement; and meanwhile, training data required by a deviation mode which is based on an agricultural specific scene is less, and the instantaneity is good, thus the robustness method provided by the invention is more applicable to application under the condition that the calculation and storage resources of the mobile equipment are limited.

Description

technical field [0001] The invention relates to the fields of intelligent information processing and agricultural information technology, in particular to a robust method for voice recognition of mobile equipment for agricultural field data collection. Background technique [0002] my country's geographical span is large, the geographical environment is complex, and the level of agricultural production and technological development is unbalanced. The application diversity and flexibility of mobile devices can help solve the front-end technical difficulties encountered in the process of agricultural informatization from the grassroots, namely Raw information collection and control issues. Speech recognition technology is an important way to solve the interaction problem of mobile devices. With the continuous deepening of agricultural modernization, the importance of speech recognition technology in the field of agricultural information is becoming more and more prominent. It i...

AI Technical Summary

Problems solved by technology

[0003] 1. The speech signal is a non-stationary signal. The commonly used noise compensation methods are based on the assumption of a linear stationary signal. The situation of considering time-varying factors and nonlinear effects is very complicated.

[0004] 2. The acoustic characteristics of continuous speech signals are very different with the speech connected before and after it, which limits the practical application of speech recognition on mobile devices with low signal-to-noise ratio

[0005] 3. The continuous speech recognition system with a large vocabulary has poor adaptability and strong dependence on the environment, and it is difficult to improve the accuracy of speech recognition in a noisy environment

[0006] It is very difficult to solve the above problems in the context of continuous speech recognition with a large vocabulary in the general field. In terms of small vocabulary applications in specific fields, the speech recognition application of mobile devices needs to adapt to the new environment of small data, while desktop speech recognition is commonly used with complex The complexity of the maximum likelihood linear regression method for parameters exceeds the computing power of mobile devices

Images