Hyperspectral imaging makes it possible to map the spatial distribution of quality-indicating attributes over biological targets. However, inconsistencies may arise between these newly available qualitymaps of the same piece of the biological target from different chemometric models. Therefore, such inconsistency of the spatial prediction of the freshness-indicating attribute of total volatile basic nitrogen (TVB-N) over pork loins (Longissimus dorsi) was investigated in this work from the perspective of the accuracy and variation of pixel-wise prediction over the top-surfaces. The partial least square regression (PLSR)-based chemometric modelling was performed on the characteristic spectra of a target extracted from regions-of-interest (ROIs) of both the whole meat and the lean part only, after spectral preprocessing and spatial filtering, and coupled with waveband selection using successive projections algorithm (SPA). Results showed that all PLSR models achieved good and stable predictions of the average TVB-N values of pork loins, despite the differences of data-preprocessing or numbers of wavebands, with R2p being in the range from 0.832 to 0.889 for Meat-ROI and an even higher range from 0.846 to 0.912 if using Lean-ROIs, even with a reduced number of wavebands, accuracy remaining high in the range of R2p from 0.722 to 0.889 for Meat-ROIs and from 0.704 to 0.900 for Lean-ROIs. In contrast, however, devastating differences emerged between the predicted TVB-N distributions over a top-surface. Results showed that excessive variation of pixel-wise predictions rendered utterly useless distribution maps resulted from a direct application of the PLSR or PLSR_SPA models from Meat or Lean ROIs, quantified by root-mean-error (RMSE) of pixel-wise predictions over 21.6 mg/100 g without PLSR, or even the worse with PLSR-SPA, shooting absurdly high about 55 mg/100 g. After appropriate spatial filtering, the excessive between-pixel variation was suppressed to below 8.5 mg/100 g in RMSE, presenting visually better distribution maps, but at a significant loss in accuracy, R2p dropping from 0.79 to a barely enough 0.58 for Meat-ROI, and from 0.86 to a failing value of 0.42 for Lean-ROI. In conclusion, though good and reliable for the average TVB-N prediction of larger surfaces such as that of pork loins, hyperspectral imaging-based prediction of TVB-N distribution is more challenging. A compromise must be made between the accuracy and visual goodness of such quality maps and a meaningful variation of the pixel-wise values, possibly with appropriate spatial filtering. And never trust a quality-indicating map, even if a visually pleasing one, without going through the evaluation of its accuracy and pixel-wise variation. Keywords: pork, TVB-N, hyperspectral imaging, prediction, evaluation, distribution, PLSR, SPA DOI: 10.25165/j.ijabe.20211404.6278 Citation: Tang Y W Y, Wang X W, Xing X Y, Li Z, Zhao M C. Prediction and evaluation method of TVB-N values distribution in pork by hyperspectral imaging. Int J Agric & Biol Eng, 2021; 14(4): 270–276.

The aim of this study is to enhance the detection accuracy of rice chlorophyll content under complex backgrounds, optimize rice production management models, and improve the efficiency and quality of grain production. To achieve this, a spectral device for rice chlorophyll content detection with integrated background classification capability was developed. The device employed a MobileNetV4-Conv-Small-based model for rice background classification, enabling the categorization of clear, muddy, and green algae-covered backgrounds. The classification results showed that the model performed best under the green algae-covered background, with all performance metrics exceeding 97%. The muddy background followed, with metrics surpassing 94%, while the clear background proved more challenging, though the metrics still exceeded 93%. By combining preprocessing techniques with convolutional neural networks (CNNs), distinct rice chlorophyll content detection models were developed for each background type. For clear backgrounds, the optimal model was FD + CNN, with R², RMSE, and RPD values of 0.975, 5.191, and 6.318, respectively. For muddy backgrounds, the optimal model was SS + CNN, with R², RMSE, and RPD values of 0.627, 18.249, and 1.638, respectively. For green algae-covered backgrounds, the SS + CNN model also achieved the best results, with R², RMSE, and RPD values of 0.719, 16.417, and 1.885, respectively. Field experiments confirmed that the device achieved a rice background classification accuracy of 94.67% and a relative error compliance rate of 84.00% for chlorophyll content detection. These results demonstrate the feasibility of integrating background classification and chlorophyll content detection for rice cultivation.