Study on wood detection based on BP neural network

Log image was acquired by X-ray real-time digital imaging system without log destruction. This paper presents the average density value of the specific spot of log is measured quickly and exactly according to log perimeter and log image information using the method of artificial neural network. According to the basic knowledge of X-ray testing technique, the method of getting high quality digital image is provided. This paper presents the improved method of image collection system, and setting method of the best parameters of equipment. This paper creates a BP network prediction model of log average density, analyses the performance of network with different structures and parameters, and presents the best parameters of BP neural networks which are used for measuring log average density. A new method of log average density fast measuring is provided.