Solid set canopy delivery system (SSCDS) are envisioned as viable alternatives to conventional spraying techniques like an axial-fan airblast sprayer in tree-fruit orchards and vineyards. While SSCDS may reduce the amount of spray drift, exposure to pesticides, carbon footprint and drive-row compaction, it needs field scale optimization in terms of emitter selection, placement and scalability. Therefore, the overall goal of this research was to customize SSCDS in model perennial specialty crops grown in the state of Washington.Different SSCDS configurations were developed using off-the-shelf emitters for a high-density tall spindle trained apple orchard and a modified vertical shoot position trained vineyard. Suitable emitter types and their within canopy placement were identified through analyzing spray performance at different canopy locations. Field spray trials indicated the requirement of emitters at multiple canopy locations for adequate spray deposition and coverage. Moreover, hollow-cone emitters facilitated spray penetration into the foliage. The optimized SSCDS provided comparable spray deposition to that of an airblast sprayer in grapevines with significantly reduced off-target drift.A pneumatic spray delivery (PSD) system was developed to address the issue of non-uniform spray application pertinent to conventional hydraulic spray delivery (HSD) based SSCDS. A reservoir sub-system was developed and integrated with the HSD to have precisely metered spray material at each of the spray locations. While the drop in operating pressure was similar in both the systems, the PSD system had uniform spray output throughout the spray row length. Such design resulted in consistently higher spray deposition and coverage at different canopy locations in a high-density apple orchard.A rapid estimation method of tree-row-volume (TRV), leaf-wall-area (LWA), canopy volume (CV) and canopy cover (CC) was developed based on unmanned aerial system (UAS) acquired aerial imagery data. Precise estimation of TRV and LWA (R2 > 0.75) was achieved. Moreover, CV and CC estimates were significantly correlated to the ground measurements (p