Study on quantitative detection of fruit tree branch bending effect based on SURF algorithm

In order to quickly and accurately evaluate the effect of fruit tree branch bending in apple orchard production process, a quantitative detection method of fruit tree branch bending effect based on SURF algorithm was proposed This study takes the three-year-old red Fuji apple tree in Yishui, Shandong Province as an example. Firstly, the image information of apple tree after branch bending operation was collected, and the binary image of apple tree branches was extracted by threshold method. Secondly, the skeleton extraction algorithm based on Euclidean distance transform is used to extract the skeleton image of fruit tree and determine the skeleton image of fruit tree trunk, extract all the pixels in the short neighborhood of the trunk, and use SURF algorithm to detect and locate the intersection of fruit tree trunk and branch. Then, the angle of fruit tree branch is calculated by using cosine theorem, and the quantitative results are obtained by comparing with the requirements of apple tree agronomic production. Finally, two groups of data were obtained by using the fruit tree branch bending detection method based on SURF algorithm and the artificial detection method respectively, and the significant difference between the two groups was tested. After testing, there is no significant difference between the two groups of data. At the same time, it takes 8.024 minutes to detect 100 fruit tree branch angle by using the quantitative detection method based on SURF algorithm, which is much less than the manual measurement time. Therefore, the quantitative detection method of fruit tree branch bending effect based on SURF algorithm can quickly and accurately measure the angle of fruit tree branch, and provide scientific basis for modern management and intelligent production of apple orchard.