Efficient information management in orchard characterization leads to more efficient agricultural processes. In this brief, a set of computational geometry methods are presented and evaluated for orchard characterization; in particular, for the estimation of canopy volume and shape in groves and orchards using a LiDAR (Light Detection And Ranging) sensor mounted on an agricultural service unit. The proposed approaches were evaluated and validated in the field, showing they are convergent in the estimation process and that they are able to estimate the crown volume for fully scanned canopies in real time; for partially observed tree crowns, accuracy decreases up to 30% (the worst case). The latter is the major contribution of this brief since it implies that the automated service unit does not need to cover all alley-ways for an accurate modeling of the orchard, thus saving valuable resources.

No abstract

No abstract

Object detection and semantic segmentation are two of the most widely adopted deep learning algorithms in agricultural applications. One of the major sources of variability in image quality acquired in the outdoors for such tasks is changing lighting condition that can alter the appearance of the objects or the contents of the entire image. While transfer learning and data augmentation to some extent reduce the need for large amount of data to train deep neural networks, the large variety of cultivars and the lack of shared datasets in agriculture makes wide-scale field deployments difficult. In this paper, we present a high throughput robust active lighting-based camera system that generates consistent images in all lighting conditions. We detail experiments that show the consistency in images quality leading to relatively fewer images to train deep neural networks for the task of object detection. We further present results from field experiment under extreme lighting conditions where images without active lighting significantly lack to provide consistent results. The experimental results show that on average, deep nets for object detection trained on consistent data required nearly four times less data to achieve similar level of accuracy. This proposed work could potentially provide pragmatic solutions to computer vision needs in agriculture.

No abstract