Chlorophyll content in plant leaves is an essential indicator of the growth condition and the fertilization management effect of naked barley crops. The soil plant analysis development (SPAD) values strongly correlate with leaf chlorophyll contents. Unmanned Aerial Vehicles (UAV) can provide an efficient way to retrieve SPAD values on a relatively large scale with a high temporal resolution. But the UAV mounted with high-cost multispectral or hyperspectral sensors may be a tremendous economic burden for smallholder farmers. To overcome this shortcoming, we investigated the potential of UAV mounted with a commercial digital camera for estimating the SPAD values of naked barley leaves. We related 21 color-based vegetation indices (VIs) calculated from UAV images acquired from two flight heights (6.0 m and 50.0 m above ground level) in four different growth stages with SPAD values. Our results indicated that vegetation extraction and naked barley ears mask could improve the correlation between image-calculated vegetation indices and SPAD values. The VIs of ‘L*,’ ‘b*,’ ‘G − B’ and ‘2G − R − B’ showed significant correlations with SPAD values of naked barley leaves at both flight heights. The validation of the regression model showed that the index of ‘G-B’ could be regarded as the most robust vegetation index for predicting the SPAD values of naked barley leaves for different images and different flight heights. Our study demonstrated that the UAV mounted with a commercial camera has great potentiality in retrieving SPAD values of naked barley leaves under unstable photography conditions. It is significant for farmers to take advantage of the cheap measurement system to monitor crops.
Robust and automated image segmentation in high-throughput image-based plant phenotyping has received considerable attention in the last decade. The possibility of this approach has not been well studied due to the time-consuming manual segmentation and lack of appropriate datasets. Segmenting images of greenhouse and open-field grown crops from the background is a challenging task linked to various factors such as complex background (presence of humans, equipment, devices, and machinery for crop management practices), environmental conditions (humidity, cloudy/sunny, fog, rain), occlusion, low-contrast and variability in crops and pose over time. This paper presents a new ubiquitous deep learning architecture ThelR547v1 (Thermal RGB 547 layers version 1) that segmented each pixel as crop or crop canopy from the background (non-crop) in real time by abstracting multi-scale contextual information with reduced memory cost. By evaluating over 37,328 augmented images (aug1: thermal RGB and RGB), our method achieves mean IoU of 0.94 and 0.87 for leaves and background and mean Bf scores of 0.93 and 0.86, respectively. ThelR547v1 has a training accuracy of 96.27%, a training loss of 0.09, a validation accuracy of 96.15%, and a validation loss of 0.10. Qualitative analysis further shows that despite the low resolution of training data, ThelR547v1 successfully distinguishes leaf/canopy pixels from complex and noisy background pixels, enabling it to be used for real-time semantic segmentation of horticultural crops.