An Accessible and Accurate Image Analysis for Root Length and Leaf Area Estimation: A Case Application to Azadirachta indica Seedlings

摘要

Root length and leaf area measurements in plants are quite cumbersome and highly error prone experimental assignments when done manually with conventional methods (such as line intersect method, inch counter method etc. for root length and geometric methods for leaf area measurement). This becomes compounded manifold when dealing with a large batch of samples. An inexpensive and accurate root length and leaf area measurement protocol was developed in this work using digital image processing procedures. The scripts were developed in MATLAB and supports user specific extensions or modifications afterwards. Three length estimators were evaluated and the best performer was selected for our use in the nursery experiment with Azadirachta indica A. Juss seedlings. Accuracy within 5% was achieved and hence served the purpose of present experiment. Image processing protocol established here was extremely fast and output was achieved in less than 8 seconds. This made the set of scripts very useful while handling a large batch of images. Further, the leaf area estimation also yielded accuracy of 2-8% of the actual area. Sources of errors and their mitigation were discussed. This technique involved commercial digital camera and flatbed scanner and hence should be more accessible compared to other complex devices. The technique was finally applied to A. indica seedlings and variation of parameters with age is reported.