| Aiming at the demand of accurate measurement of forest structure parameters in the fields of forest carbon sequestration measurement,forest resource informatics and ecosystem research,this paper proposes to improve the fuzzy c-means clustering(FCM)algorithm to improve the effectiveness and accuracy of forest structure parameters measurement by taking tree height as an example and using close range photogrammetry technology.As the basic unit of forest stand,standing trees are irregular distributed objects in three-dimensional space,and their shape is affected by wind speed and other environmental factors,which makes it difficult for high-precision ranging instruments to be directly applied,resulting in low measurement accuracy and large difference in multiple measurement results.In order to effectively alleviate the above problems,this paper proposes to use the mobile phone equipped with fisheye lens to obtain hemispherical images of trees,and establish the tree height measurement model according to the equidistant projection principle of fisheye lens,so as to avoid the interference of wind speed,reduce the influence of human factors,and improve the effectiveness of single measurement.In order to improve the measurement accuracy,the scaramuzza model is introduced to calibrate the fish eye camera,and the parameters of the tree height measurement model are obtained,this paper improves the FCM algorithm to segment the tree image,uses the super-pixel algorithm to preprocess the image to reduce the computational complexity,and introduces Salp Swarm Algorithmic to optimize the initial clustering center of FCM algorithm,so as to solve the sensitivity of FCM algorithm to the initial clustering center,so as to realize the accurate extraction of tree extreme points,and takes the extreme points as the input of tree height calculation model to realize the calculation of tree height.In the accuracy verification experiment of the tree height measurement model,the distance between the corner points in the black-and-white checkerboard calibration board is calculated.Through 50 times of measurement experiments,the results show that the maximum relative error of the measurement model is 1.274%,the minimum relative error is 0.323%,and the average relative error is 0.823%,which proves the effectiveness of the tree height measurement model.The field measurement results of tree height show that the proposed method is effective The maximum relative error of tree height measurement method is 3.050%,the minimum relative error is 0.500%,and the average relative error is 1.446%.The accuracy of the method is higher than that of the widely used ultrasonic distance measurement method,which can meet the requirement that the error of tree height measurement is not more than 5%in the national forest resources survey.The improved FCM algorithm proposed in this paper can provide technical support for accurate measurement of forest structure parameters. |
