Previous PageTable Of ContentsNext Page

Rapid Image Analysis for counting engorged pollen grains of rice

K.M. Fox1,2, T.C. Farrell 1,2,3, and R.L. Williams1,2

1YAI, PMB, Yanco NSW.
2
CRC for Sustainable Rice Production, Yanco, NSW.

3School of Land and Food, The University of Queensland, Brisbane, Qld.

ABSTRACT

Low temperatures during the early pollen microspore stage of rice (Oryza sativa) causes considerable yield loss to the Australian Rice Industry. A key measure of a variety’s response to cold damage is the number of engorged pollen grains it has produced. A reliable image analysis system has been developed to count the number of engorged pollen grains in rice anthers. Manual counts were compared with the image analysis system, resulting in a strong correlation (r2=0.99).

KEY WORDS

Rice, image analysis, low temperature, pollen grains.

INTRODUCTION

Low temperatures during the early pollen microspore stage of rice (Oryza sativa) causes considerable yield loss to the Australian Rice Industry. It is estimated that the average total cost due to cold damage (resulting in spikelet sterility) is $A20M (Farrell et al, 2000). The differential varietal response to low temperatures is being investigated to identify the mechanisms involved in cold induced sterility. A major focus of the research is to identify cold tolerant varieties at flowering by counting engorged pollen grains. Research has indicated that the total number of engorged pollen grains is an important factor for determining the cold tolerance of varieties (Nishiyama, 1996; Bechar et al, 1997). A reliable image analysis system has been developed to count the number of engorged pollen grains in rice anthers.

MATERIALS AND METHODS

Rice anthers were soaked in 0.8g of Sigma® cellulase and 12mls of sodium acetate buffer (pH=5), placed in a waterbath at 37°C and left overnight. Four anthers (two each from separate spikelets) were selected at random from individual varieties. The anthers were manually dissected (in a viewing dish) and the pollen grains stained with potassium iodide to detect the presence of starch. Those pollen grains that stained black contained starch and were therefore classified as engorged. Pollen grains were then counted using image analysis.

The image analysis system comprised a Pentium 300Mhz computer linked to a JVC video camera, which was attached to a dissecting microscope (Figure 1). Images of the pollen grains were captured (magnification 16x) and saved as a digital image using Matrox Rainbow Runner. Pollen grains outside the field of view of the digital image were recorded separately. The image was viewed in Sigma ScanPro Version 4® for counting. The picture was saved and modified by converting to grey scale and maximising the contrast. A superior image was obtained and accurate counting of the pollen grains proceeded.

Figure 1. The image analysis system in operation.

RESULTS AND DISCUSSION

To determine the accuracy of the image analysis system, manual counts were completed on many images encompassing a wide range of varieties and pollen numbers (0-2200). The image analysis system was found to be very accurate with a strong correlation (r2 = 0.99) between the computer and manual counts.

This system has enabled a large number of pollen grains to be processed efficiently and accurately. Many varieties including cold tolerant and sensitive varieties have been tested using the image analysis system. HSC 55 (a cold tolerant variety from Hungary) had a high average pollen number of 1800, compared to Doongara (a sensitive variety from Australia) which had an average pollen number of 750, following exposure to air temperatures of 27°C/13°C (day/night) during early pollen microspore (Figure 2A and 2B).

Figure 2. HSC 55, a cold tolerant variety from Hungary (A), compared to Doongara, a sensitive variety from Australia (B) following exposure to air temperatures of 27°C/13°C (day/night).

CONCLUSION

The image analysis system that was developed to count rice pollen numbers was found to be a very efficient and accurate method for determining the number of engorged pollen grains. Compared to manual counting the image analysis system resulted in time savings of 30 minutes per image. In the future this system could be successfully used to calculate pollen numbers of other crops.

REFERENCES

1. Bechar, A.; Gan-mor, S.; Vaknin, Y.; Shmulevich, I.; Ronen, B. and Eisikowitch, D. 1997. An image-analysis technique for accurate counting of pollen on stigmas. New Phytologist, 137, 639-643.

2. Farrell, T.C.; Williams, R.L. and Fukai, S. 2000. The cost of low temperature to the NSW Rice Industry. Proceedings 10th Australian Agronomy Conference, Hobart.

3. Nishiyama, I. 1996. Strategies for the research to overcome cool weather damage in rice plants. Proceedings of the 2nd Asian Crop Science Conference. 246-251.

Previous PageTop Of PageNext Page