In fruit production, critical crop management decisions are guided by bloom intensity, i.e., the number of flowers present in an orchard. Despite its importance, bloom intensity is still typically estimated by means of human visual inspection. Existing automated computer vision systems for flower identification are based on hand-engineered techniques that work only under specific conditions and with limited performance. This letter proposes an automated technique for flower identification that is robust to uncontrolled environments and applicable to different flower species. Our method relies on an end-to-end residual convolutional neural network (CNN) that represents the state-of-the-art in semantic segmentation. To enhance its sensitivity to flowers, we fine-tune this network using a single dataset of apple flower images. Since CNNs tend to produce coarse segmentations, we employ a refinement method to better distinguish between individual flower instances. Without any preprocessing or dataset-specific training, experimental results on images of apple, peach, and pear flowers, acquired under different conditions demonstrate the robustness and broad applicability of our method.

Significance: Trees’ branches grow against the pull of gravity and toward light. Although gravity and light perception have been studied in model species, much is unknown about how trees detect and respond to these signals. Here, we report the identification of a gene ( WEEP ) that controls lateral branch orientations and is directly or indirectly required for gravity responses in trees. Loss or reduction of WEEP expression produced branches that grow outward and downward and did not exhibit normal gravitropism responses when displaced. WEEP is conserved throughout the plant kingdom and may be involved in gravity perception. WEEP may also be a valuable target for breeding or engineering trees with improved shapes for agricultural and landscaping applications.

Prunus persica (peach) trees carrying the “Pillar” or “Broomy” trait (br) have vertically oriented branches caused by loss-of-function mutations in a gene called TILLER ANGLE CONTROL 1 (TAC1). TAC1 encodes a protein in the IGT gene family that includes LAZY1 and DEEPER ROOTING 1 (DRO1), which regulate lateral branch and root orientations, respectively. Here we found that some of the native TAC1 alleles in the hexaploid plum species Prunus domestica, which has a naturally more upright stature, contained a variable length trinucleotide repeat within the same exon 3 region previously found to be disrupted in pillar peach trees. RNAi silencing of TAC1 in plum resulted in trees with severely vertical branch orientations similar to those in pillar peaches but with an even narrower profile. In contrast, PpeTAC1 overexpression in plum led to trees with wider branch angles and more horizontal branch orientations. Pillar peach trees and transgenic plum lines exhibited pleiotropic phenotypes, including differences in trunk and branch diameter, stem growth, and twisting branch phenotypes. Expression profiling of pillar peach trees revealed differential expression of numerous genes associated with biotic and abiotic stress, hormone responses, plastids, reactive oxygen, secondary, and cell wall metabolism. Collectively, the data provide important clues for understanding TAC1 function and show that alteration of TAC1 expression may have broad applicability to agricultural and ornamental tree industries.