Fishery analysis is critical for ensuring the sustainability of marine species and supporting the livelihoods of millions who rely on fishing for food and income. Commercial catch analysis provides essential insights for stock assessments and management by estimating fishing effort and its ecological impacts. Accurate fish size data is vital for fish stock assessment modeling and to understand fishing impacts, but manual size sampling and annotation of daily landings are impractical and error-prone. In this paper, we leverage a unique data set of fish images and measurements from 2022 to 2024 to solve fish length distribution sampling and prediction tasks. We demonstrate that extracting key information from images can enhance predictive performance, often surpassing more complex methods. Furthermore, species-specific segmentation models outperformed those trained on mixed classes, underscoring the importance of tailored segmentation in achieving accurate predictions. This work highlights the potential of automated systems to improve fish population stock assessments, a critical and complex challenge in management and conservation efforts.