The Application Of Technology Image Pcessing And Statistical Approach In Automatic Tree Rings Analysis System

With the social development and progress of technology, the technology of imageprocessing will be used in more fields of nature science. Tree ring analysis can provideuseful data like life length and yearly growth of a tree for dendrochronologists. Theanalysis can help them to researches on past ecologies, climate, environmental changesand archaeology. Because counting number of rings and measuring widths aretime-consuming and hard tasks, an automatic system is recently demanded.The target of this paper is to complete the analysis system which can accomplishthe two above-mentioned functions. And the paper analyzes the difficulties in the designprocess of the system: scars which come across the rings and other huge scars; noise onthe cross section of the tree; self-conditioned algorithm. Mean-wile the research of thispaper is focus on the algorithm of rings detection, noise reduction and measurement toresolve the difficulties. In the measurement, statistic approach is introduced to avoid theinfluence of scars and raise the rate of accurate; label algorithm is used to get the area ofeach connect domains to reduce the noise; self-conditioned binalization method isintroduced to get the outline of the rings. Based on the requirement of normal ringsanalysis system, this paper makes a brief introduction the procedure of image processand rings measurement. This paper chooses appropriate method and parameter to getsatisfying results by experiment. At the same time, it proves the rationality of themethod in this paper by experiment.The method in this paper accomplishes the two necessary functions and mostlyresolves the difficulties in the research. The experiments show that the margin of errorin the measurement of the number of the rings is about±10%and the proposed methodhas an advantage in the accuracy when comparing with traditional methods.