Abstract. Image analysis is a useful tool for seed monitoring during germination. An important advantage of the image analysis consists in easy determination of the dimensional changes in time without any manipulation with the germinated seeds. Many researches focus on the final germination phase and the image analysis is used for automatic detection of the visible radicle protrusion from the seed coat. However, an interesting behavior can be found in the whole germination process. The aim of this paper consists in development of an automatic method for observation and detection the image processes connected with seed germination. The setup consists of a photographic camera, a stand, an illumination LED panel and a glass vessel for the germination test. The camera is controlled via a computer and images are stored in the hard disk. The area of the seeds and their shape parameters were determined by a special program. The development of image moments is presented in our work. It was found that image moments can be useful tools for automatic detection of the primary shoot during the final germination phase as well as during the imbibition process.Keywords: imbibition, water movement, image moments, image analysis.
IntroductionThe seed germination is a critical process for plant propagations. Germination starts by a quick water uptake and it is terminated with an elongation of the embryonic axis [1]. The final germination stage is usually represented by penetration of the radicle through the structures surrounding the embryo. This stage is often called visible germination [2].Image analysis is a useful tool for seed monitoring during germination. Shouche et al.[3] tested several shape parameters of Indian wheat seeds as a tool for identification of different varieties. The authors stated that the use of over 45 different parameters enabled them to distinguish between 15 samples of Indian wheat varieties. The moment analysis played a key role in their work. Kuhl and Giardina [4] introduced in their work a method for shape analysis based on elliptic Fourier descriptors. This method was widely used for shape analysis of plant organs by many researchers [5][6][7]. Mebatsion et al. [8] developed an algorithm to classify cereal grains. The authors tested their algorithm with barley, oat, rye and two varieties of wheat. The results show that the method can be used for seed classification. The classification accuracies were almost 100 %.Dell'Aquila [9] dealt with the application of the digital imaging information technology to seed germination testing. In his review article, he presents several parameters, which can be monitored during germination. The presented parameters are, e.g.: area, perimeter, length, width, roundness, calculated with the formula: perimeter 2 / 4π Area, and aspect, calculated with the ratio between the longer axis and the shorter axis of the ellipse equivalent to the seed area. The author noted one particular example of roundness factor utilization: a seed with a circular shape has the roundness f...