Biological imaging is one of the major tools used to decipher the functioning of meristems in plant morphogenesis. Extracting quantitative information from microscopy images typically requires complex image processing and analysis pipelines, designed by combining elementary operations such as signal enhancement, object segmentation, and object measurements. Moreover, integrating data through the processing of large image datasets is required to model variability and ensure robust analyses. Existing image analysis software solutions fall into two categories that come with their respective limitations. On the one hand, dedicated libraries such as ITK provide collections of basic components that can be assembled into possibly complex image processing pipelines. This requires a strong computer programming expertise. On the other hand, embedded software such as Fiji provide user-friendly graphical interfaces that allow end users with little or no programming expertise to perform a wide range of predefined image processing operations. The possibilities for automation and batch processing are however often limited, and generally do not eliminate the need for the user to do some programming. We address these limitations by proposing a new software, called BIP, for batch processing and analysis of biological images. BIP integrates many standard algorithms and specific algorithms developed in the framework of our research projects for quantifying images of plant cells and tissues. BIP fills-in an empty niche in the ecosystem of bioimaging software, by relying on a simple yet powerful command line interface. Furthermore, integrating BIP within frameworks for high-throughput, distributed computing is straightforward. BIP allows users to easily specify sets of images to be processed and to readily chain basic operations into complex analysis pipelines. BIP offers transparent support for images of arbitrary dimensions (2D to 5D) and numerical types (integer and floating numbers, signed and unsigned, from 8 to 32 bits), thus breaking the compatibility barriers often encountered between implementations of different algorithms in existing software. This performance is achieved thanks to a single, unified templated data structure to represent multi-dimensional images and a generic design pattern for the automatic detection of the numerical data type in input images. In addition, these design principles allow minimizing the number of implementations of each algorithm, thus easying code maintenance, evolution, and optimization. Here, we illustrate through several examples the benefits and potential of BIP for processing and quantifying image datasets of plant cells and tissues, including 3D time-lapse confocal images of A. thaliana shoot apical meristems.