Digital steganography has a long history, starting to be developed in the 90s until now. The main aspects of early steganography are security, imperceptibility, and payload. Security is an essential aspect of steganography, where its development competes with the steganalysis method. The development of steganographic methods is not only based on statistics but also involves adaptive methods, machine learning, artificial intelligence, generative adversarial networks (GAN), convolution neural networks (CNN), coverless techniques, etc. Several new methods add robustness as an important aspect. This makes the development of steganographic methods more varied and generally focuses the goal of the method on one or two aspects. Nowadays, steganography articles, especially surveys, rarely explicitly classify steganography based on its goals. So this survey contributes to classifying steganographic research based on goals and its assessments. This paper also reviews the use of assessment tools in depth because it is closely related to the goal of steganography. Challenges or issues, methods, developments, and popular datasets are also reviewed. Steganalysis literature was also added to improve the quality of the study and analysis. Finally, this survey provides discussion, investigation, critical analysis, and explicit summary so novice researchers can more easily understand image steganography development.