The advancement in imaging devices has permitted high-throughput study of plant traits also termed as plant phenotyping. In this context, among the different imaging modalities employed in plant phenotyping, chlorophyll fluorescence imaging provides accurate characterisation of crucial plant traits such as photosynthetic efficiency, water use efficiency and nutrients use efficiency to name a few. However, chlorophyll fluorescence imaging is complex and expensive in contrast to widely adopted visible (RGB) images. This limits the reliable study of the aforementioned plant traits. To relieve this limitation, chlorophyll fluorescence image reconstruction algorithms from widely available visible images are a prime solution. Thus in this paper, we propose a domain translation-based approach using cycle consistency loss for chlorophyll fluorescence image reconstruction from RGB images. The experimental results on the maize dataset demonstrate the efficiency of our approach with respect to this novel application.